/[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 178 by ph10, Wed Jun 13 08:44:34 2007 UTC revision 227 by ph10, Tue Aug 21 15:00:15 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 62  used by pcretest. DEBUG is not defined w Line 65  used by pcretest. DEBUG is not defined w
65    
66  #define SETBIT(a,b) a[b/8] |= (1 << (b%8))  #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         *
# Line 120  static const short int escapes[] = { Line 130  static const short int escapes[] = {
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 */-ESC_H,     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,-ESC_V, -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,
# Line 130  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 203  static const char *error_texts[] = { Line 234  static const char *error_texts[] = {
234    "missing ) after comment",    "missing ) after comment",
235    "parentheses nested too deeply",  /** DEAD **/    "parentheses nested too deeply",  /** DEAD **/
236    /* 20 */    /* 20 */
237    "regular expression too large",    "regular expression is too large",
238    "failed to get memory",    "failed to get memory",
239    "unmatched parentheses",    "unmatched parentheses",
240    "internal error: code overflow",    "internal error: code overflow",
# Line 239  static const char *error_texts[] = { Line 270  static const char *error_texts[] = {
270    "subpattern name is too long (maximum " XSTRING(MAX_NAME_SIZE) " characters)",    "subpattern name is too long (maximum " XSTRING(MAX_NAME_SIZE) " characters)",
271    "too many named subpatterns (maximum " XSTRING(MAX_NAME_COUNT) ")",    "too many named subpatterns (maximum " XSTRING(MAX_NAME_COUNT) ")",
272    /* 50 */    /* 50 */
273    "repeated subpattern is too long",    "repeated subpattern is too long",    /** DEAD **/
274    "octal value is greater than \\377 (not in UTF-8 mode)",    "octal value is greater than \\377 (not in UTF-8 mode)",
275    "internal error: overran compiling workspace",    "internal error: overran compiling workspace",
276    "internal error: previously-checked referenced subpattern not found",    "internal error: previously-checked referenced subpattern not found",
# Line 248  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 379  static const unsigned char ebcdic_charta Line 414  static const unsigned char ebcdic_charta
414  /* Definition to allow mutual recursion */  /* Definition to allow mutual recursion */
415    
416  static BOOL  static BOOL
417    compile_regex(int, int, uschar **, const uschar **, int *, BOOL, BOOL, int,    compile_regex(int, int, uschar **, const uschar **, int *, BOOL, BOOL, int,
418      int *, int *, branch_chain *, compile_data *, int *);      int *, int *, branch_chain *, compile_data *, int *);
419    
420    
# Line 405  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 490  else Line 525  else
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 501  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 530  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 625  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 701  if (c == '{') Line 747  if (c == '{')
747      *negptr = TRUE;      *negptr = TRUE;
748      ptr++;      ptr++;
749      }      }
750    for (i = 0; i < sizeof(name) - 1; i++)    for (i = 0; i < (int)sizeof(name) - 1; i++)
751      {      {
752      c = *(++ptr);      c = *(++ptr);
753      if (c == 0) goto ERROR_RETURN;      if (c == 0) goto ERROR_RETURN;
# Line 904  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 931  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 1059  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 1148  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 1256  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_TYPEPOSSTAR:
1321          case OP_TYPEPOSPLUS:
1322          case OP_TYPEPOSQUERY:
1323          if (code[1] == OP_PROP || code[1] == OP_NOTPROP) code += 2;
1324          break;
1325    
1326          case OP_TYPEUPTO:
1327          case OP_TYPEMINUPTO:
1328          case OP_TYPEEXACT:
1329          case OP_TYPEPOSUPTO:
1330          if (code[3] == OP_PROP || code[3] == OP_NOTPROP) code += 2;
1331          break;
1332          }
1333    
1334        /* Add in the fixed length from the table */
1335    
1336      code += _pcre_OP_lengths[c];      code += _pcre_OP_lengths[c];
1337    
1338      /* In UTF-8 mode, opcodes that are followed by a character may be followed by
1339      a multi-byte character. The length in the table is a minimum, so we have to
1340      arrange to skip the extra bytes. */
1341    
1342  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
1343      if (utf8) switch(c)      if (utf8) switch(c)
1344        {        {
# Line 1320  for (;;) Line 1396  for (;;)
1396    
1397    if (c == OP_XCLASS) code += GET(code, 1);    if (c == OP_XCLASS) code += GET(code, 1);
1398    
1399    /* 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
1400    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
1401    The length in the table is a minimum, so we have to arrange to skip the extra    two bytes of parameters. */
   bytes. */  
1402    
1403    else    else
1404      {      {
1405        switch(c)
1406          {
1407          case OP_TYPESTAR:
1408          case OP_TYPEMINSTAR:
1409          case OP_TYPEPLUS:
1410          case OP_TYPEMINPLUS:
1411          case OP_TYPEQUERY:
1412          case OP_TYPEMINQUERY:
1413          case OP_TYPEPOSSTAR:
1414          case OP_TYPEPOSPLUS:
1415          case OP_TYPEPOSQUERY:
1416          if (code[1] == OP_PROP || code[1] == OP_NOTPROP) code += 2;
1417          break;
1418    
1419          case OP_TYPEPOSUPTO:
1420          case OP_TYPEUPTO:
1421          case OP_TYPEMINUPTO:
1422          case OP_TYPEEXACT:
1423          if (code[3] == OP_PROP || code[3] == OP_NOTPROP) code += 2;
1424          break;
1425          }
1426    
1427        /* Add in the fixed length from the table */
1428    
1429      code += _pcre_OP_lengths[c];      code += _pcre_OP_lengths[c];
1430    
1431        /* In UTF-8 mode, opcodes that are followed by a character may be followed
1432        by a multi-byte character. The length in the table is a minimum, so we have
1433        to arrange to skip the extra bytes. */
1434    
1435  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
1436      if (utf8) switch(c)      if (utf8) switch(c)
1437        {        {
# Line 1399  for (code = first_significant_code(code Line 1503  for (code = first_significant_code(code
1503    
1504    /* For other groups, scan the branches. */    /* For other groups, scan the branches. */
1505    
1506    if (c == OP_BRA || c == OP_CBRA || c == OP_ONCE)    if (c == OP_BRA || c == OP_CBRA || c == OP_ONCE || c == OP_COND)
1507      {      {
1508      BOOL empty_branch;      BOOL empty_branch;
1509      if (GET(code, 1) == 0) return TRUE;    /* Hit unclosed bracket */      if (GET(code, 1) == 0) return TRUE;    /* Hit unclosed bracket */
# Line 1423  for (code = first_significant_code(code Line 1527  for (code = first_significant_code(code
1527    
1528    switch (c)    switch (c)
1529      {      {
1530      /* Check for quantifiers after a class */      /* Check for quantifiers after a class. XCLASS is used for classes that
1531        cannot be represented just by a bit map. This includes negated single
1532        high-valued characters. The length in _pcre_OP_lengths[] is zero; the
1533        actual length is stored in the compiled code, so we must update "code"
1534        here. */
1535    
1536  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
1537      case OP_XCLASS:      case OP_XCLASS:
1538      ccode = code + GET(code, 1);      ccode = code += GET(code, 1);
1539      goto CHECK_CLASS_REPEAT;      goto CHECK_CLASS_REPEAT;
1540  #endif  #endif
1541    
# Line 1489  for (code = first_significant_code(code Line 1597  for (code = first_significant_code(code
1597      case OP_TYPEEXACT:      case OP_TYPEEXACT:
1598      return FALSE;      return FALSE;
1599    
1600        /* These are going to continue, as they may be empty, but we have to
1601        fudge the length for the \p and \P cases. */
1602    
1603        case OP_TYPESTAR:
1604        case OP_TYPEMINSTAR:
1605        case OP_TYPEPOSSTAR:
1606        case OP_TYPEQUERY:
1607        case OP_TYPEMINQUERY:
1608        case OP_TYPEPOSQUERY:
1609        if (code[1] == OP_PROP || code[1] == OP_NOTPROP) code += 2;
1610        break;
1611    
1612        /* Same for these */
1613    
1614        case OP_TYPEUPTO:
1615        case OP_TYPEMINUPTO:
1616        case OP_TYPEPOSUPTO:
1617        if (code[3] == OP_PROP || code[3] == OP_NOTPROP) code += 2;
1618        break;
1619    
1620      /* End of branch */      /* End of branch */
1621    
1622      case OP_KET:      case OP_KET:
# Line 1651  adjust_recurse(uschar *group, int adjust Line 1779  adjust_recurse(uschar *group, int adjust
1779    uschar *save_hwm)    uschar *save_hwm)
1780  {  {
1781  uschar *ptr = group;  uschar *ptr = group;
1782    
1783  while ((ptr = (uschar *)find_recurse(ptr, utf8)) != NULL)  while ((ptr = (uschar *)find_recurse(ptr, utf8)) != NULL)
1784    {    {
1785    int offset;    int offset;
# Line 1949  if (next >= 0) switch(op_code) Line 2078  if (next >= 0) switch(op_code)
2078    case OP_NOT_WORDCHAR:    case OP_NOT_WORDCHAR:
2079    return next <= 127 && (cd->ctypes[next] & ctype_word) != 0;    return next <= 127 && (cd->ctypes[next] & ctype_word) != 0;
2080    
2081      case OP_HSPACE:
2082      case OP_NOT_HSPACE:
2083      switch(next)
2084        {
2085        case 0x09:
2086        case 0x20:
2087        case 0xa0:
2088        case 0x1680:
2089        case 0x180e:
2090        case 0x2000:
2091        case 0x2001:
2092        case 0x2002:
2093        case 0x2003:
2094        case 0x2004:
2095        case 0x2005:
2096        case 0x2006:
2097        case 0x2007:
2098        case 0x2008:
2099        case 0x2009:
2100        case 0x200A:
2101        case 0x202f:
2102        case 0x205f:
2103        case 0x3000:
2104        return op_code != OP_HSPACE;
2105        default:
2106        return op_code == OP_HSPACE;
2107        }
2108    
2109      case OP_VSPACE:
2110      case OP_NOT_VSPACE:
2111      switch(next)
2112        {
2113        case 0x0a:
2114        case 0x0b:
2115        case 0x0c:
2116        case 0x0d:
2117        case 0x85:
2118        case 0x2028:
2119        case 0x2029:
2120        return op_code != OP_VSPACE;
2121        default:
2122        return op_code == OP_VSPACE;
2123        }
2124    
2125    default:    default:
2126    return FALSE;    return FALSE;
2127    }    }
# Line 1983  switch(op_code) Line 2156  switch(op_code)
2156      case ESC_W:      case ESC_W:
2157      return item <= 127 && (cd->ctypes[item] & ctype_word) != 0;      return item <= 127 && (cd->ctypes[item] & ctype_word) != 0;
2158    
2159        case ESC_h:
2160        case ESC_H:
2161        switch(item)
2162          {
2163          case 0x09:
2164          case 0x20:
2165          case 0xa0:
2166          case 0x1680:
2167          case 0x180e:
2168          case 0x2000:
2169          case 0x2001:
2170          case 0x2002:
2171          case 0x2003:
2172          case 0x2004:
2173          case 0x2005:
2174          case 0x2006:
2175          case 0x2007:
2176          case 0x2008:
2177          case 0x2009:
2178          case 0x200A:
2179          case 0x202f:
2180          case 0x205f:
2181          case 0x3000:
2182          return -next != ESC_h;
2183          default:
2184          return -next == ESC_h;
2185          }
2186    
2187        case ESC_v:
2188        case ESC_V:
2189        switch(item)
2190          {
2191          case 0x0a:
2192          case 0x0b:
2193          case 0x0c:
2194          case 0x0d:
2195          case 0x85:
2196          case 0x2028:
2197          case 0x2029:
2198          return -next != ESC_v;
2199          default:
2200          return -next == ESC_v;
2201          }
2202    
2203      default:      default:
2204      return FALSE;      return FALSE;
2205      }      }
2206    
2207    case OP_DIGIT:    case OP_DIGIT:
2208    return next == -ESC_D || next == -ESC_s || next == -ESC_W;    return next == -ESC_D || next == -ESC_s || next == -ESC_W ||
2209             next == -ESC_h || next == -ESC_v;
2210    
2211    case OP_NOT_DIGIT:    case OP_NOT_DIGIT:
2212    return next == -ESC_d;    return next == -ESC_d;
# Line 1997  switch(op_code) Line 2215  switch(op_code)
2215    return next == -ESC_S || next == -ESC_d || next == -ESC_w;    return next == -ESC_S || next == -ESC_d || next == -ESC_w;
2216    
2217    case OP_NOT_WHITESPACE:    case OP_NOT_WHITESPACE:
2218    return next == -ESC_s;    return next == -ESC_s || next == -ESC_h || next == -ESC_v;
2219    
2220      case OP_HSPACE:
2221      return next == -ESC_S || next == -ESC_H || next == -ESC_d || next == -ESC_w;
2222    
2223      case OP_NOT_HSPACE:
2224      return next == -ESC_h;
2225    
2226      /* Can't have \S in here because VT matches \S (Perl anomaly) */
2227      case OP_VSPACE:
2228      return next == -ESC_V || next == -ESC_d || next == -ESC_w;
2229    
2230      case OP_NOT_VSPACE:
2231      return next == -ESC_v;
2232    
2233    case OP_WORDCHAR:    case OP_WORDCHAR:
2234    return next == -ESC_W || next == -ESC_s;    return next == -ESC_W || next == -ESC_s || next == -ESC_h || next == -ESC_v;
2235    
2236    case OP_NOT_WORDCHAR:    case OP_NOT_WORDCHAR:
2237    return next == -ESC_w || next == -ESC_d;    return next == -ESC_w || next == -ESC_d;
# Line 2115  for (;; ptr++) Line 2346  for (;; ptr++)
2346    BOOL possessive_quantifier;    BOOL possessive_quantifier;
2347    BOOL is_quantifier;    BOOL is_quantifier;
2348    BOOL is_recurse;    BOOL is_recurse;
2349    BOOL reset_bracount;    BOOL reset_bracount;
2350    int class_charcount;    int class_charcount;
2351    int class_lastchar;    int class_lastchar;
2352    int newoptions;    int newoptions;
# Line 2153  for (;; ptr++) Line 2384  for (;; ptr++)
2384      */      */
2385    
2386      if (code < last_code) code = last_code;      if (code < last_code) code = last_code;
2387    
2388        /* Paranoid check for integer overflow */
2389    
2390        if (OFLOW_MAX - *lengthptr < code - last_code)
2391          {
2392          *errorcodeptr = ERR20;
2393          goto FAILED;
2394          }
2395    
2396      *lengthptr += code - last_code;      *lengthptr += code - last_code;
2397      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));
2398    
# Line 2265  for (;; ptr++) Line 2505  for (;; ptr++)
2505      *ptrptr = ptr;      *ptrptr = ptr;
2506      if (lengthptr != NULL)      if (lengthptr != NULL)
2507        {        {
2508          if (OFLOW_MAX - *lengthptr < code - last_code)
2509            {
2510            *errorcodeptr = ERR20;
2511            goto FAILED;
2512            }
2513        *lengthptr += code - last_code;   /* To include callout length */        *lengthptr += code - last_code;   /* To include callout length */
2514        DPRINTF((">> end branch\n"));        DPRINTF((">> end branch\n"));
2515        }        }
# Line 2327  for (;; ptr++) Line 2572  for (;; ptr++)
2572        goto FAILED;        goto FAILED;
2573        }        }
2574    
2575      /* 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,
2576        if the first few characters (either before or after ^) are \Q\E or \E we
2577        skip them too. This makes for compatibility with Perl. */
2578    
2579      if ((c = *(++ptr)) == '^')      negate_class = FALSE;
2580        for (;;)
2581        {        {
       negate_class = TRUE;  
2582        c = *(++ptr);        c = *(++ptr);
2583        }        if (c == '\\')
2584      else          {
2585        {          if (ptr[1] == 'E') ptr++;
2586        negate_class = FALSE;            else if (strncmp((const char *)ptr+1, "Q\\E", 3) == 0) ptr += 3;
2587                else break;
2588            }
2589          else if (!negate_class && c == '^')
2590            negate_class = TRUE;
2591          else break;
2592        }        }
2593    
2594      /* 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
# Line 2477  for (;; ptr++) Line 2729  for (;; ptr++)
2729        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
2730        case. Inside a class (and only there) it is treated as backspace.        case. Inside a class (and only there) it is treated as backspace.
2731        Elsewhere it marks a word boundary. Other escapes have preset maps ready        Elsewhere it marks a word boundary. Other escapes have preset maps ready
2732        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
2733        character in them, so set class_charcount bigger than one. */        character in them, so set class_charcount bigger than one. */
2734    
2735        if (c == '\\')        if (c == '\\')
# Line 2497  for (;; ptr++) Line 2749  for (;; ptr++)
2749            else inescq = TRUE;            else inescq = TRUE;
2750            continue;            continue;
2751            }            }
2752            else if (-c == ESC_E) continue;  /* Ignore orphan \E */
2753    
2754          if (c < 0)          if (c < 0)
2755            {            {
# Line 2535  for (;; ptr++) Line 2788  for (;; ptr++)
2788    
2789              case ESC_E: /* Perl ignores an orphan \E */              case ESC_E: /* Perl ignores an orphan \E */
2790              continue;              continue;
2791    
2792              default:    /* Not recognized; fall through */              default:    /* Not recognized; fall through */
2793              break;      /* Need "default" setting to stop compiler warning. */              break;      /* Need "default" setting to stop compiler warning. */
2794              }              }
# Line 2544  for (;; ptr++) Line 2797  for (;; ptr++)
2797    
2798            else if (c == -ESC_d || c == -ESC_D || c == -ESC_w ||            else if (c == -ESC_d || c == -ESC_D || c == -ESC_w ||
2799                     c == -ESC_W || c == -ESC_s || c == -ESC_S) continue;                     c == -ESC_W || c == -ESC_s || c == -ESC_S) continue;
2800    
2801            /* We need to deal with \H, \h, \V, and \v in both phases because            /* We need to deal with \H, \h, \V, and \v in both phases because
2802            they use extra memory. */            they use extra memory. */
2803    
2804            if (-c == ESC_h)            if (-c == ESC_h)
2805              {              {
2806              SETBIT(classbits, 0x09); /* VT */              SETBIT(classbits, 0x09); /* VT */
2807              SETBIT(classbits, 0x20); /* SPACE */              SETBIT(classbits, 0x20); /* SPACE */
2808              SETBIT(classbits, 0xa0); /* NSBP */              SETBIT(classbits, 0xa0); /* NSBP */
2809  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
2810              if (utf8)              if (utf8)
2811                {                {
2812                class_utf8 = TRUE;                class_utf8 = TRUE;
2813                *class_utf8data++ = XCL_SINGLE;                *class_utf8data++ = XCL_SINGLE;
2814                class_utf8data += _pcre_ord2utf8(0x1680, class_utf8data);                class_utf8data += _pcre_ord2utf8(0x1680, class_utf8data);
2815                *class_utf8data++ = XCL_SINGLE;                *class_utf8data++ = XCL_SINGLE;
2816                class_utf8data += _pcre_ord2utf8(0x180e, class_utf8data);                class_utf8data += _pcre_ord2utf8(0x180e, class_utf8data);
2817                *class_utf8data++ = XCL_RANGE;                *class_utf8data++ = XCL_RANGE;
2818                class_utf8data += _pcre_ord2utf8(0x2000, class_utf8data);                class_utf8data += _pcre_ord2utf8(0x2000, class_utf8data);
2819                class_utf8data += _pcre_ord2utf8(0x200A, class_utf8data);                class_utf8data += _pcre_ord2utf8(0x200A, class_utf8data);
2820                *class_utf8data++ = XCL_SINGLE;                *class_utf8data++ = XCL_SINGLE;
2821                class_utf8data += _pcre_ord2utf8(0x202f, class_utf8data);                class_utf8data += _pcre_ord2utf8(0x202f, class_utf8data);
2822                *class_utf8data++ = XCL_SINGLE;                *class_utf8data++ = XCL_SINGLE;
2823                class_utf8data += _pcre_ord2utf8(0x205f, class_utf8data);                class_utf8data += _pcre_ord2utf8(0x205f, class_utf8data);
2824                *class_utf8data++ = XCL_SINGLE;                *class_utf8data++ = XCL_SINGLE;
2825                class_utf8data += _pcre_ord2utf8(0x3000, class_utf8data);                class_utf8data += _pcre_ord2utf8(0x3000, class_utf8data);
2826                }                }
2827  #endif  #endif
2828              continue;              continue;
2829              }              }
2830    
2831            if (-c == ESC_H)            if (-c == ESC_H)
2832              {              {
# Line 2581  for (;; ptr++) Line 2834  for (;; ptr++)
2834                {                {
2835                int x = 0xff;                int x = 0xff;
2836                switch (c)                switch (c)
2837                  {                  {
2838                  case 0x09/8: x ^= 1 << (0x09%8); break;                  case 0x09/8: x ^= 1 << (0x09%8); break;
2839                  case 0x20/8: x ^= 1 << (0x20%8); break;                  case 0x20/8: x ^= 1 << (0x20%8); break;
2840                  case 0xa0/8: x ^= 1 << (0xa0%8); break;                  case 0xa0/8: x ^= 1 << (0xa0%8); break;
2841                  default: break;                  default: break;
2842                  }                  }
2843                classbits[c] |= x;                classbits[c] |= x;
2844                }                }
2845    
2846  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
2847              if (utf8)              if (utf8)
2848                {                {
2849                class_utf8 = TRUE;                class_utf8 = TRUE;
2850                *class_utf8data++ = XCL_RANGE;                *class_utf8data++ = XCL_RANGE;
2851                class_utf8data += _pcre_ord2utf8(0x0100, class_utf8data);                class_utf8data += _pcre_ord2utf8(0x0100, class_utf8data);
2852                class_utf8data += _pcre_ord2utf8(0x167f, class_utf8data);                class_utf8data += _pcre_ord2utf8(0x167f, class_utf8data);
2853                *class_utf8data++ = XCL_RANGE;                *class_utf8data++ = XCL_RANGE;
2854                class_utf8data += _pcre_ord2utf8(0x1681, class_utf8data);                class_utf8data += _pcre_ord2utf8(0x1681, class_utf8data);
2855                class_utf8data += _pcre_ord2utf8(0x180d, class_utf8data);                class_utf8data += _pcre_ord2utf8(0x180d, class_utf8data);
2856                *class_utf8data++ = XCL_RANGE;                *class_utf8data++ = XCL_RANGE;
2857                class_utf8data += _pcre_ord2utf8(0x180f, class_utf8data);                class_utf8data += _pcre_ord2utf8(0x180f, class_utf8data);
2858                class_utf8data += _pcre_ord2utf8(0x1fff, class_utf8data);                class_utf8data += _pcre_ord2utf8(0x1fff, class_utf8data);
2859                *class_utf8data++ = XCL_RANGE;                *class_utf8data++ = XCL_RANGE;
2860                class_utf8data += _pcre_ord2utf8(0x200B, class_utf8data);                class_utf8data += _pcre_ord2utf8(0x200B, class_utf8data);
2861                class_utf8data += _pcre_ord2utf8(0x202e, class_utf8data);                class_utf8data += _pcre_ord2utf8(0x202e, class_utf8data);
2862                *class_utf8data++ = XCL_RANGE;                *class_utf8data++ = XCL_RANGE;
2863                class_utf8data += _pcre_ord2utf8(0x2030, class_utf8data);                class_utf8data += _pcre_ord2utf8(0x2030, class_utf8data);
2864                class_utf8data += _pcre_ord2utf8(0x205e, class_utf8data);                class_utf8data += _pcre_ord2utf8(0x205e, class_utf8data);
2865                *class_utf8data++ = XCL_RANGE;                *class_utf8data++ = XCL_RANGE;
2866                class_utf8data += _pcre_ord2utf8(0x2060, class_utf8data);                class_utf8data += _pcre_ord2utf8(0x2060, class_utf8data);
2867                class_utf8data += _pcre_ord2utf8(0x2fff, class_utf8data);                class_utf8data += _pcre_ord2utf8(0x2fff, class_utf8data);
2868                *class_utf8data++ = XCL_RANGE;                *class_utf8data++ = XCL_RANGE;
2869                class_utf8data += _pcre_ord2utf8(0x3001, class_utf8data);                class_utf8data += _pcre_ord2utf8(0x3001, class_utf8data);
2870                class_utf8data += _pcre_ord2utf8(0x7fffffff, class_utf8data);                class_utf8data += _pcre_ord2utf8(0x7fffffff, class_utf8data);
2871                }                }
2872  #endif  #endif
2873              continue;              continue;
2874              }              }
2875    
2876            if (-c == ESC_v)            if (-c == ESC_v)
2877              {              {
2878              SETBIT(classbits, 0x0a); /* LF */              SETBIT(classbits, 0x0a); /* LF */
2879              SETBIT(classbits, 0x0b); /* VT */              SETBIT(classbits, 0x0b); /* VT */
2880              SETBIT(classbits, 0x0c); /* FF */              SETBIT(classbits, 0x0c); /* FF */
2881              SETBIT(classbits, 0x0d); /* CR */              SETBIT(classbits, 0x0d); /* CR */
2882              SETBIT(classbits, 0x85); /* NEL */              SETBIT(classbits, 0x85); /* NEL */
2883  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
2884              if (utf8)              if (utf8)
2885                {                {
2886                class_utf8 = TRUE;                class_utf8 = TRUE;
2887                *class_utf8data++ = XCL_RANGE;                *class_utf8data++ = XCL_RANGE;
2888                class_utf8data += _pcre_ord2utf8(0x2028, class_utf8data);                class_utf8data += _pcre_ord2utf8(0x2028, class_utf8data);
2889                class_utf8data += _pcre_ord2utf8(0x2029, class_utf8data);                class_utf8data += _pcre_ord2utf8(0x2029, class_utf8data);
2890                }                }
2891  #endif  #endif
2892              continue;              continue;
2893              }              }
2894    
2895            if (-c == ESC_V)            if (-c == ESC_V)
2896              {              {
# Line 2645  for (;; ptr++) Line 2898  for (;; ptr++)
2898                {                {
2899                int x = 0xff;                int x = 0xff;
2900                switch (c)                switch (c)
2901                  {                  {
2902                  case 0x0a/8: x ^= 1 << (0x0a%8);                  case 0x0a/8: x ^= 1 << (0x0a%8);
2903                               x ^= 1 << (0x0b%8);                               x ^= 1 << (0x0b%8);
2904                               x ^= 1 << (0x0c%8);                               x ^= 1 << (0x0c%8);
2905                               x ^= 1 << (0x0d%8);                               x ^= 1 << (0x0d%8);
2906                               break;                               break;
2907                  case 0x85/8: x ^= 1 << (0x85%8); break;                  case 0x85/8: x ^= 1 << (0x85%8); break;
2908                  default: break;                  default: break;
2909                  }                  }
2910                classbits[c] |= x;                classbits[c] |= x;
2911                }                }
2912    
2913  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
2914              if (utf8)              if (utf8)
2915                {                {
2916                class_utf8 = TRUE;                class_utf8 = TRUE;
2917                *class_utf8data++ = XCL_RANGE;                *class_utf8data++ = XCL_RANGE;
2918                class_utf8data += _pcre_ord2utf8(0x0100, class_utf8data);                class_utf8data += _pcre_ord2utf8(0x0100, class_utf8data);
2919                class_utf8data += _pcre_ord2utf8(0x2027, class_utf8data);                class_utf8data += _pcre_ord2utf8(0x2027, class_utf8data);
2920                *class_utf8data++ = XCL_RANGE;                *class_utf8data++ = XCL_RANGE;
2921                class_utf8data += _pcre_ord2utf8(0x2029, class_utf8data);                class_utf8data += _pcre_ord2utf8(0x2029, class_utf8data);
2922                class_utf8data += _pcre_ord2utf8(0x7fffffff, class_utf8data);                class_utf8data += _pcre_ord2utf8(0x7fffffff, class_utf8data);
2923                }                }
2924  #endif  #endif
2925              continue;              continue;
2926              }              }
2927    
2928            /* We need to deal with \P and \p in both phases. */            /* We need to deal with \P and \p in both phases. */
2929    
# Line 2812  for (;; ptr++) Line 3065  for (;; ptr++)
3065              unsigned int origd = d;              unsigned int origd = d;
3066              while (get_othercase_range(&cc, origd, &occ, &ocd))              while (get_othercase_range(&cc, origd, &occ, &ocd))
3067                {                {
3068                if (occ >= (unsigned int)c &&                if (occ >= (unsigned int)c &&
3069                    ocd <= (unsigned int)d)                    ocd <= (unsigned int)d)
3070                  continue;                          /* Skip embedded ranges */                  continue;                          /* Skip embedded ranges */
3071    
3072                if (occ < (unsigned int)c  &&                if (occ < (unsigned int)c  &&
3073                    ocd >= (unsigned int)c - 1)      /* Extend the basic range */                    ocd >= (unsigned int)c - 1)      /* Extend the basic range */
3074                  {                                  /* if there is overlap,   */                  {                                  /* if there is overlap,   */
3075                  c = occ;                           /* noting that if occ < c */                  c = occ;                           /* noting that if occ < c */
3076                  continue;                          /* we can't have ocd > d  */                  continue;                          /* we can't have ocd > d  */
3077                  }                                  /* because a subrange is  */                  }                                  /* because a subrange is  */
3078                if (ocd > (unsigned int)d &&                if (ocd > (unsigned int)d &&
3079                    occ <= (unsigned int)d + 1)      /* always shorter than    */                    occ <= (unsigned int)d + 1)      /* always shorter than    */
3080                  {                                  /* the basic range.       */                  {                                  /* the basic range.       */
3081                  d = ocd;                  d = ocd;
# Line 2943  for (;; ptr++) Line 3196  for (;; ptr++)
3196        goto FAILED;        goto FAILED;
3197        }        }
3198    
3199        /* Remember whether \r or \n are in this class */
3200    
3201        if (negate_class)
3202          {
3203          if ((classbits[1] & 0x24) != 0x24) cd->external_options |= PCRE_HASCRORLF;
3204          }
3205        else
3206          {
3207          if ((classbits[1] & 0x24) != 0) cd->external_options |= PCRE_HASCRORLF;
3208          }
3209    
3210      /* 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
3211      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
3212      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
3213      because OP_NOT and the related opcodes like OP_NOTSTAR operate on      optimize.
3214      single-bytes only. This is an historical hangover. Maybe one day we can  
3215      tidy these opcodes to handle multi-byte characters.      In UTF-8 mode, we can optimize the negative case only if there were no
3216        characters >= 128 because OP_NOT and the related opcodes like OP_NOTSTAR
3217        operate on single-bytes only. This is an historical hangover. Maybe one day
3218        we can tidy these opcodes to handle multi-byte characters.
3219    
3220      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
3221      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 2958  for (;; ptr++) Line 3225  for (;; ptr++)
3225      reqbyte, save the previous value for reinstating. */      reqbyte, save the previous value for reinstating. */
3226    
3227  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
3228      if (class_charcount == 1 &&      if (class_charcount == 1 && !class_utf8 &&
3229            (!utf8 ||        (!utf8 || !negate_class || class_lastchar < 128))
           (!class_utf8 && (!negate_class || class_lastchar < 128))))  
   
3230  #else  #else
3231      if (class_charcount == 1)      if (class_charcount == 1)
3232  #endif  #endif
# Line 3419  for (;; ptr++) Line 3684  for (;; ptr++)
3684          goto FAILED;          goto FAILED;
3685          }          }
3686    
       /* This is a paranoid check to stop integer overflow later on */  
   
       if (len > MAX_DUPLENGTH)  
         {  
         *errorcodeptr = ERR50;  
         goto FAILED;  
         }  
   
3687        /* 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
3688        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
3689        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 3515  for (;; ptr++) Line 3772  for (;; ptr++)
3772          if (repeat_min > 1)          if (repeat_min > 1)
3773            {            {
3774            /* 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
3775            just adjust the length as if we had. */            just adjust the length as if we had. Do some paranoid checks for
3776              potential integer overflow. */
3777    
3778            if (lengthptr != NULL)            if (lengthptr != NULL)
3779              *lengthptr += (repeat_min - 1)*length_prevgroup;              {
3780                int delta = (repeat_min - 1)*length_prevgroup;
3781                if ((double)(repeat_min - 1)*(double)length_prevgroup >
3782                                                                (double)INT_MAX ||
3783                    OFLOW_MAX - *lengthptr < delta)
3784                  {
3785                  *errorcodeptr = ERR20;
3786                  goto FAILED;
3787                  }
3788                *lengthptr += delta;
3789                }
3790    
3791            /* This is compiling for real */            /* This is compiling for real */
3792    
# Line 3556  for (;; ptr++) Line 3824  for (;; ptr++)
3824          /* 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
3825          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
3826          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
3827          add 2 + 2*LINKSIZE to allow for the nesting that occurs. */          add 2 + 2*LINKSIZE to allow for the nesting that occurs. Do some
3828            paranoid checks to avoid integer overflow. */
3829    
3830          if (lengthptr != NULL && repeat_max > 0)          if (lengthptr != NULL && repeat_max > 0)
3831            *lengthptr += repeat_max * (length_prevgroup + 1 + 2 + 2*LINK_SIZE) -            {
3832              2 - 2*LINK_SIZE;  /* Last one doesn't nest */            int delta = repeat_max * (length_prevgroup + 1 + 2 + 2*LINK_SIZE) -
3833                          2 - 2*LINK_SIZE;   /* Last one doesn't nest */
3834              if ((double)repeat_max *
3835                    (double)(length_prevgroup + 1 + 2 + 2*LINK_SIZE)
3836                      > (double)INT_MAX ||
3837                  OFLOW_MAX - *lengthptr < delta)
3838                {
3839                *errorcodeptr = ERR20;
3840                goto FAILED;
3841                }
3842              *lengthptr += delta;
3843              }
3844    
3845          /* This is compiling for real */          /* This is compiling for real */
3846    
# Line 3712  for (;; ptr++) Line 3992  for (;; ptr++)
3992      /* ===================================================================*/      /* ===================================================================*/
3993      /* Start of nested parenthesized sub-expression, or comment or lookahead or      /* Start of nested parenthesized sub-expression, or comment or lookahead or
3994      lookbehind or option setting or condition or all the other extended      lookbehind or option setting or condition or all the other extended
3995      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. */  
3996    
3997      case '(':      case '(':
3998      newoptions = options;      newoptions = options;
3999      skipbytes = 0;      skipbytes = 0;
4000      bravalue = OP_CBRA;      bravalue = OP_CBRA;
4001      save_hwm = cd->hwm;      save_hwm = cd->hwm;
4002      reset_bracount = FALSE;      reset_bracount = FALSE;
4003    
4004        /* First deal with various "verbs" that can be introduced by '*'. */
4005    
4006      if (*(++ptr) == '?')      if (*(++ptr) == '*' && (cd->ctypes[ptr[1]] & ctype_letter) != 0)
4007          {
4008          int i, namelen;
4009          const uschar *name = ++ptr;
4010          previous = NULL;
4011          while ((cd->ctypes[*++ptr] & ctype_letter) != 0);
4012          if (*ptr == ':')
4013            {
4014            *errorcodeptr = ERR59;   /* Not supported */
4015            goto FAILED;
4016            }
4017          if (*ptr != ')')
4018            {
4019            *errorcodeptr = ERR60;
4020            goto FAILED;
4021            }
4022          namelen = ptr - name;
4023          for (i = 0; i < verbcount; i++)
4024            {
4025            if (namelen == verbs[i].len &&
4026                strncmp((char *)name, verbs[i].name, namelen) == 0)
4027              {
4028              *code = verbs[i].op;
4029              if (*code++ == OP_ACCEPT) cd->had_accept = TRUE;
4030              break;
4031              }
4032            }
4033          if (i < verbcount) continue;
4034          *errorcodeptr = ERR60;
4035          goto FAILED;
4036          }
4037    
4038        /* Deal with the extended parentheses; all are introduced by '?', and the
4039        appearance of any of them means that this is not a capturing group. */
4040    
4041        else if (*ptr == '?')
4042        {        {
4043        int i, set, unset, namelen;        int i, set, unset, namelen;
4044        int *optset;        int *optset;
# Line 3746  for (;; ptr++) Line 4061  for (;; ptr++)
4061          /* ------------------------------------------------------------ */          /* ------------------------------------------------------------ */
4062          case '|':                 /* Reset capture count for each branch */          case '|':                 /* Reset capture count for each branch */
4063          reset_bracount = TRUE;          reset_bracount = TRUE;
4064          /* Fall through */          /* Fall through */
4065    
4066          /* ------------------------------------------------------------ */          /* ------------------------------------------------------------ */
4067          case ':':                 /* Non-capturing bracket */          case ':':                 /* Non-capturing bracket */
# Line 3965  for (;; ptr++) Line 4280  for (;; ptr++)
4280    
4281          /* ------------------------------------------------------------ */          /* ------------------------------------------------------------ */
4282          case '!':                 /* Negative lookahead */          case '!':                 /* Negative lookahead */
         bravalue = OP_ASSERT_NOT;  
4283          ptr++;          ptr++;
4284            if (*ptr == ')')          /* Optimize (?!) */
4285              {
4286              *code++ = OP_FAIL;
4287              previous = NULL;
4288              continue;
4289              }
4290            bravalue = OP_ASSERT_NOT;
4291          break;          break;
4292    
4293    
# Line 4447  for (;; ptr++) Line 4768  for (;; ptr++)
4768           errorcodeptr,                 /* Where to put an error message */           errorcodeptr,                 /* Where to put an error message */
4769           (bravalue == OP_ASSERTBACK ||           (bravalue == OP_ASSERTBACK ||
4770            bravalue == OP_ASSERTBACK_NOT), /* TRUE if back assert */            bravalue == OP_ASSERTBACK_NOT), /* TRUE if back assert */
4771           reset_bracount,               /* True if (?| group */           reset_bracount,               /* True if (?| group */
4772           skipbytes,                    /* Skip over bracket number */           skipbytes,                    /* Skip over bracket number */
4773           &subfirstbyte,                /* For possible first char */           &subfirstbyte,                /* For possible first char */
4774           &subreqbyte,                  /* For possible last char */           &subreqbyte,                  /* For possible last char */
# Line 4515  for (;; ptr++) Line 4836  for (;; ptr++)
4836        goto FAILED;        goto FAILED;
4837        }        }
4838    
4839      /* 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,
4840      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
4841      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
4842      a quantifier. */      duplicated by a quantifier.*/
4843    
4844      if (lengthptr != NULL)      if (lengthptr != NULL)
4845        {        {
4846          if (OFLOW_MAX - *lengthptr < length_prevgroup - 2 - 2*LINK_SIZE)
4847            {
4848            *errorcodeptr = ERR20;
4849            goto FAILED;
4850            }
4851        *lengthptr += length_prevgroup - 2 - 2*LINK_SIZE;        *lengthptr += length_prevgroup - 2 - 2*LINK_SIZE;
4852        code++;        *code++ = OP_BRA;
4853        PUTINC(code, 0, 1 + LINK_SIZE);        PUTINC(code, 0, 1 + LINK_SIZE);
4854        *code++ = OP_KET;        *code++ = OP_KET;
4855        PUTINC(code, 0, 1 + LINK_SIZE);        PUTINC(code, 0, 1 + LINK_SIZE);
4856          break;    /* No need to waste time with special character handling */
4857        }        }
4858    
4859      /* Otherwise update the main code pointer to the end of the group. */      /* Otherwise update the main code pointer to the end of the group. */
4860    
4861      else code = tempcode;      code = tempcode;
4862    
4863      /* For a DEFINE group, required and first character settings are not      /* For a DEFINE group, required and first character settings are not
4864      relevant. */      relevant. */
# Line 4735  for (;; ptr++) Line 5062  for (;; ptr++)
5062      *code++ = ((options & PCRE_CASELESS) != 0)? OP_CHARNC : OP_CHAR;      *code++ = ((options & PCRE_CASELESS) != 0)? OP_CHARNC : OP_CHAR;
5063      for (c = 0; c < mclength; c++) *code++ = mcbuffer[c];      for (c = 0; c < mclength; c++) *code++ = mcbuffer[c];
5064    
5065        /* Remember if \r or \n were seen */
5066    
5067        if (mcbuffer[0] == '\r' || mcbuffer[0] == '\n')
5068          cd->external_options |= PCRE_HASCRORLF;
5069    
5070      /* Set the first and required bytes appropriately. If no previous first      /* Set the first and required bytes appropriately. If no previous first
5071      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.
5072      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 4807  Arguments: Line 5139  Arguments:
5139    ptrptr         -> the address of the current pattern pointer    ptrptr         -> the address of the current pattern pointer
5140    errorcodeptr   -> pointer to error code variable    errorcodeptr   -> pointer to error code variable
5141    lookbehind     TRUE if this is a lookbehind assertion    lookbehind     TRUE if this is a lookbehind assertion
5142    reset_bracount TRUE to reset the count for each branch    reset_bracount TRUE to reset the count for each branch
5143    skipbytes      skip this many bytes at start (for brackets and OP_COND)    skipbytes      skip this many bytes at start (for brackets and OP_COND)
5144    firstbyteptr   place to put the first required character, or a negative number    firstbyteptr   place to put the first required character, or a negative number
5145    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 4821  Returns:         TRUE on success Line 5153  Returns:         TRUE on success
5153    
5154  static BOOL  static BOOL
5155  compile_regex(int options, int oldims, uschar **codeptr, const uschar **ptrptr,  compile_regex(int options, int oldims, uschar **codeptr, const uschar **ptrptr,
5156    int *errorcodeptr, BOOL lookbehind, BOOL reset_bracount, int skipbytes,    int *errorcodeptr, BOOL lookbehind, BOOL reset_bracount, int skipbytes,
5157    int *firstbyteptr, int *reqbyteptr, branch_chain *bcptr, compile_data *cd,    int *firstbyteptr, int *reqbyteptr, branch_chain *bcptr, compile_data *cd,
5158    int *lengthptr)    int *lengthptr)
5159  {  {
5160  const uschar *ptr = *ptrptr;  const uschar *ptr = *ptrptr;
# Line 4867  orig_bracount = max_bracount = cd->braco Line 5199  orig_bracount = max_bracount = cd->braco
5199  for (;;)  for (;;)
5200    {    {
5201    /* For a (?| group, reset the capturing bracket count so that each branch    /* For a (?| group, reset the capturing bracket count so that each branch
5202    uses the same numbers. */    uses the same numbers. */
5203    
5204    if (reset_bracount) cd->bracount = orig_bracount;    if (reset_bracount) cd->bracount = orig_bracount;
5205    
5206    /* Handle a change of ims options at the start of the branch */    /* Handle a change of ims options at the start of the branch */
5207    
5208    if ((options & PCRE_IMS) != oldims)    if ((options & PCRE_IMS) != oldims)
# Line 4899  for (;;) Line 5231  for (;;)
5231      *ptrptr = ptr;      *ptrptr = ptr;
5232      return FALSE;      return FALSE;
5233      }      }
5234    
5235    /* Keep the highest bracket count in case (?| was used and some branch    /* Keep the highest bracket count in case (?| was used and some branch
5236    has fewer than the rest. */    has fewer than the rest. */
5237    
5238    if (cd->bracount > max_bracount) max_bracount = cd->bracount;    if (cd->bracount > max_bracount) max_bracount = cd->bracount;
5239    
5240    /* 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 5006  for (;;) Line 5338  for (;;)
5338        *code++ = oldims;        *code++ = oldims;
5339        length += 2;        length += 2;
5340        }        }
5341    
5342      /* Retain the highest bracket number, in case resetting was used. */      /* Retain the highest bracket number, in case resetting was used. */
5343    
5344      cd->bracount = max_bracount;      cd->bracount = max_bracount;
5345    
5346      /* Set values to pass back */      /* Set values to pass back */
# Line 5017  for (;;) Line 5349  for (;;)
5349      *ptrptr = ptr;      *ptrptr = ptr;
5350      *firstbyteptr = firstbyte;      *firstbyteptr = firstbyte;
5351      *reqbyteptr = reqbyte;      *reqbyteptr = reqbyte;
5352      if (lengthptr != NULL) *lengthptr += length;      if (lengthptr != NULL)
5353          {
5354          if (OFLOW_MAX - *lengthptr < length)
5355            {
5356            *errorcodeptr = ERR20;
5357            return FALSE;
5358            }
5359          *lengthptr += length;
5360          }
5361      return TRUE;      return TRUE;
5362      }      }
5363    
# Line 5326  real_pcre *re; Line 5666  real_pcre *re;
5666  int length = 1;  /* For final END opcode */  int length = 1;  /* For final END opcode */
5667  int firstbyte, reqbyte, newline;  int firstbyte, reqbyte, newline;
5668  int errorcode = 0;  int errorcode = 0;
5669    int skipatstart = 0;
5670  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
5671  BOOL utf8;  BOOL utf8;
5672  #endif  #endif
# Line 5404  cd->fcc = tables + fcc_offset; Line 5745  cd->fcc = tables + fcc_offset;
5745  cd->cbits = tables + cbits_offset;  cd->cbits = tables + cbits_offset;
5746  cd->ctypes = tables + ctypes_offset;  cd->ctypes = tables + ctypes_offset;
5747    
5748    /* Check for newline settings at the start of the pattern, and remember the
5749    offset for later. */
5750    
5751    if (ptr[0] == '(' && ptr[1] == '*')
5752      {
5753      int newnl = 0;
5754      if (strncmp((char *)(ptr+2), "CR)", 3) == 0)
5755        { skipatstart = 5; newnl = PCRE_NEWLINE_CR; }
5756      else if (strncmp((char *)(ptr+2), "LF)", 3)  == 0)
5757        { skipatstart = 5; newnl = PCRE_NEWLINE_LF; }
5758      else if (strncmp((char *)(ptr+2), "CRLF)", 5)  == 0)
5759        { skipatstart = 7; newnl = PCRE_NEWLINE_CR + PCRE_NEWLINE_LF; }
5760      else if (strncmp((char *)(ptr+2), "ANY)", 4) == 0)
5761        { skipatstart = 6; newnl = PCRE_NEWLINE_ANY; }
5762      else if (strncmp((char *)(ptr+2), "ANYCRLF)", 8)  == 0)
5763        { skipatstart = 10; newnl = PCRE_NEWLINE_ANYCRLF; }
5764      if (skipatstart > 0)
5765        options = (options & ~PCRE_NEWLINE_BITS) | newnl;
5766      }
5767    
5768  /* Handle different types of newline. The three bits give seven cases. The  /* Handle different types of newline. The three bits give seven cases. The
5769  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
5770  "anycrlf". */  "anycrlf". */
5771    
5772  switch (options & (PCRE_NEWLINE_CRLF | PCRE_NEWLINE_ANY))  switch (options & PCRE_NEWLINE_BITS)
5773    {    {
5774    case 0: newline = NEWLINE; break;   /* Compile-time default */    case 0: newline = NEWLINE; break;   /* Build-time default */
5775    case PCRE_NEWLINE_CR: newline = '\r'; break;    case PCRE_NEWLINE_CR: newline = '\r'; break;
5776    case PCRE_NEWLINE_LF: newline = '\n'; break;    case PCRE_NEWLINE_LF: newline = '\n'; break;
5777    case PCRE_NEWLINE_CR+    case PCRE_NEWLINE_CR+
# Line 5482  been put into the cd block so that they Line 5843  been put into the cd block so that they
5843  found within the regex right at the beginning. Bringing initial option settings  found within the regex right at the beginning. Bringing initial option settings
5844  outside can help speed up starting point checks. */  outside can help speed up starting point checks. */
5845    
5846    ptr += skipatstart;
5847  code = cworkspace;  code = cworkspace;
5848  *code = OP_BRA;  *code = OP_BRA;
5849  (void)compile_regex(cd->external_options, cd->external_options & PCRE_IMS,  (void)compile_regex(cd->external_options, cd->external_options & PCRE_IMS,
5850    &code, &ptr, &errorcode, FALSE, FALSE, 0, &firstbyte, &reqbyte, NULL, cd,    &code, &ptr, &errorcode, FALSE, FALSE, 0, &firstbyte, &reqbyte, NULL, cd,
5851    &length);    &length);
5852  if (errorcode != 0) goto PCRE_EARLY_ERROR_RETURN;  if (errorcode != 0) goto PCRE_EARLY_ERROR_RETURN;
5853    
# Line 5545  cd->start_code = codestart; Line 5907  cd->start_code = codestart;
5907  cd->hwm = cworkspace;  cd->hwm = cworkspace;
5908  cd->req_varyopt = 0;  cd->req_varyopt = 0;
5909  cd->nopartial = FALSE;  cd->nopartial = FALSE;
5910    cd->had_accept = FALSE;
5911    
5912  /* Set up a starting, non-extracting bracket, then compile the expression. On  /* Set up a starting, non-extracting bracket, then compile the expression. On
5913  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
5914  of the function here. */  of the function here. */
5915    
5916  ptr = (const uschar *)pattern;  ptr = (const uschar *)pattern + skipatstart;
5917  code = (uschar *)codestart;  code = (uschar *)codestart;
5918  *code = OP_BRA;  *code = OP_BRA;
5919  (void)compile_regex(re->options, re->options & PCRE_IMS, &code, &ptr,  (void)compile_regex(re->options, re->options & PCRE_IMS, &code, &ptr,
# Line 5559  re->top_bracket = cd->bracount; Line 5922  re->top_bracket = cd->bracount;
5922  re->top_backref = cd->top_backref;  re->top_backref = cd->top_backref;
5923    
5924  if (cd->nopartial) re->options |= PCRE_NOPARTIAL;  if (cd->nopartial) re->options |= PCRE_NOPARTIAL;
5925    if (cd->had_accept) reqbyte = -1;   /* Must disable after (*ACCEPT) */
5926    
5927  /* 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. */
5928    
# Line 5657  case when building a production library. Line 6021  case when building a production library.
6021  printf("Length = %d top_bracket = %d top_backref = %d\n",  printf("Length = %d top_bracket = %d top_backref = %d\n",
6022    length, re->top_bracket, re->top_backref);    length, re->top_bracket, re->top_backref);
6023    
6024  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 " : "");  
   }  
6025    
6026  if ((re->options & PCRE_FIRSTSET) != 0)  if ((re->options & PCRE_FIRSTSET) != 0)
6027    {    {

Legend:
Removed from v.178  
changed lines
  Added in v.227

  ViewVC Help
Powered by ViewVC 1.1.5