/[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 1065 by chpe, Tue Oct 16 15:54:16 2012 UTC revision 1133 by chpe, Thu Oct 18 18:35:30 2012 UTC
# Line 84  static int Line 84  static int
84      const pcre_uint32 *, unsigned int);      const pcre_uint32 *, unsigned int);
85    
86  static BOOL  static BOOL
87    compile_regex(int, pcre_uchar **, const pcre_uchar **, int *, BOOL, BOOL,    compile_regex(int, pcre_uchar **, const pcre_uchar **, int *, BOOL, BOOL, int, int,
88      int, int, int *, int *, branch_chain *, compile_data *, int *);      pcre_uint32 *, pcre_int32 *, pcre_uint32 *, pcre_int32 *, branch_chain *,
89        compile_data *, int *);
90    
91    
92    
# Line 121  overrun before it actually does run off Line 122  overrun before it actually does run off
122    
123  /* Private flags added to firstchar and reqchar. */  /* Private flags added to firstchar and reqchar. */
124    
125  #define REQ_CASELESS   0x10000000l      /* Indicates caselessness */  #define REQ_CASELESS    (1 << 0)        /* Indicates caselessness */
126  #define REQ_VARY       0x20000000l      /* Reqchar followed non-literal item */  #define REQ_VARY        (1 << 1)        /* Reqchar followed non-literal item */
127  #define REQ_MASK       (REQ_CASELESS | REQ_VARY)  /* Negative values for the firstchar and reqchar flags */
128    #define REQ_UNSET       (-2)
129    #define REQ_NONE        (-1)
130    
131  /* Repeated character flags. */  /* Repeated character flags. */
132    
# Line 645  static const pcre_uint8 ebcdic_chartab[] Line 648  static const pcre_uint8 ebcdic_chartab[]
648    
649    
650    
   
651  /*************************************************  /*************************************************
652  *            Find an error text                  *  *            Find an error text                  *
653  *************************************************/  *************************************************/
# Line 1201  escape sequence. Line 1203  escape sequence.
1203  Argument:  Argument:
1204    ptrptr         points to the pattern position pointer    ptrptr         points to the pattern position pointer
1205    negptr         points to a boolean that is set TRUE for negation else FALSE    negptr         points to a boolean that is set TRUE for negation else FALSE
1206    dptr           points to an int that is set to the detailed property value    ptypeptr       points to an unsigned int that is set to the type value
1207      pdataptr       points to an unsigned int that is set to the detailed property value
1208    errorcodeptr   points to the error code variable    errorcodeptr   points to the error code variable
1209    
1210  Returns:         type value from ucp_type_table, or -1 for an invalid type  Returns:         TRUE if the type value was found, or FALSE for an invalid type
1211  */  */
1212    
1213  static int  static BOOL
1214  get_ucp(const pcre_uchar **ptrptr, BOOL *negptr, int *dptr, int *errorcodeptr)  get_ucp(const pcre_uchar **ptrptr, BOOL *negptr, unsigned int *ptypeptr,
1215      unsigned int *pdataptr, int *errorcodeptr)
1216  {  {
1217  pcre_uchar c;  pcre_uchar c;
1218  int i, bot, top;  int i, bot, top;
# Line 1263  while (bot < top) Line 1267  while (bot < top)
1267    r = STRCMP_UC_C8(name, PRIV(utt_names) + PRIV(utt)[i].name_offset);    r = STRCMP_UC_C8(name, PRIV(utt_names) + PRIV(utt)[i].name_offset);
1268    if (r == 0)    if (r == 0)
1269      {      {
1270      *dptr = PRIV(utt)[i].value;      *ptypeptr = PRIV(utt)[i].type;
1271      return PRIV(utt)[i].type;      *pdataptr = PRIV(utt)[i].value;
1272        return TRUE;
1273      }      }
1274    if (r > 0) bot = i + 1; else top = i;    if (r > 0) bot = i + 1; else top = i;
1275    }    }
1276    
1277  *errorcodeptr = ERR47;  *errorcodeptr = ERR47;
1278  *ptrptr = ptr;  *ptrptr = ptr;
1279  return -1;  return FALSE;
1280    
1281  ERROR_RETURN:  ERROR_RETURN:
1282  *errorcodeptr = ERR46;  *errorcodeptr = ERR46;
1283  *ptrptr = ptr;  *ptrptr = ptr;
1284  return -1;  return FALSE;
1285  }  }
1286  #endif  #endif
1287    
# Line 1311  int max = -1; Line 1316  int max = -1;
1316  /* Read the minimum value and do a paranoid check: a negative value indicates  /* Read the minimum value and do a paranoid check: a negative value indicates
1317  an integer overflow. */  an integer overflow. */
1318    
1319  while (IS_DIGIT(*p)) min = min * 10 + *p++ - CHAR_0;  while (IS_DIGIT(*p)) min = min * 10 + (int)(*p++ - CHAR_0);
1320  if (min < 0 || min > 65535)  if (min < 0 || min > 65535)
1321    {    {
1322    *errorcodeptr = ERR5;    *errorcodeptr = ERR5;
# Line 1326  if (*p == CHAR_RIGHT_CURLY_BRACKET) max Line 1331  if (*p == CHAR_RIGHT_CURLY_BRACKET) max
1331    if (*(++p) != CHAR_RIGHT_CURLY_BRACKET)    if (*(++p) != CHAR_RIGHT_CURLY_BRACKET)
1332      {      {
1333      max = 0;      max = 0;
1334      while(IS_DIGIT(*p)) max = max * 10 + *p++ - CHAR_0;      while(IS_DIGIT(*p)) max = max * 10 + (int)(*p++ - CHAR_0);
1335      if (max < 0 || max > 65535)      if (max < 0 || max > 65535)
1336        {        {
1337        *errorcodeptr = ERR5;        *errorcodeptr = ERR5;
# Line 1457  if (ptr[0] == CHAR_LEFT_PARENTHESIS) Line 1462  if (ptr[0] == CHAR_LEFT_PARENTHESIS)
1462      if ((*ptr == CHAR_LESS_THAN_SIGN && ptr[1] != CHAR_EXCLAMATION_MARK &&      if ((*ptr == CHAR_LESS_THAN_SIGN && ptr[1] != CHAR_EXCLAMATION_MARK &&
1463          ptr[1] != CHAR_EQUALS_SIGN) || *ptr == CHAR_APOSTROPHE)          ptr[1] != CHAR_EQUALS_SIGN) || *ptr == CHAR_APOSTROPHE)
1464        {        {
1465        int term;        pcre_uchar term;
1466        const pcre_uchar *thisname;        const pcre_uchar *thisname;
1467        *count += 1;        *count += 1;
1468        if (name == NULL && *count == lorn) return *count;        if (name == NULL && *count == lorn) return *count;
# Line 1465  if (ptr[0] == CHAR_LEFT_PARENTHESIS) Line 1470  if (ptr[0] == CHAR_LEFT_PARENTHESIS)
1470        if (term == CHAR_LESS_THAN_SIGN) term = CHAR_GREATER_THAN_SIGN;        if (term == CHAR_LESS_THAN_SIGN) term = CHAR_GREATER_THAN_SIGN;
1471        thisname = ptr;        thisname = ptr;
1472        while (*ptr != term) ptr++;        while (*ptr != term) ptr++;
1473        if (name != NULL && lorn == ptr - thisname &&        if (name != NULL && lorn == (int)(ptr - thisname) &&
1474            STRNCMP_UC_UC(name, thisname, lorn) == 0)            STRNCMP_UC_UC(name, thisname, (unsigned int)lorn) == 0)
1475          return *count;          return *count;
1476        term++;        term++;
1477        }        }
# Line 1744  for (;;) Line 1749  for (;;)
1749    {    {
1750    int d;    int d;
1751    pcre_uchar *ce, *cs;    pcre_uchar *ce, *cs;
1752    register int op = *cc;    register pcre_uchar op = *cc;
1753    
1754    switch (op)    switch (op)
1755      {      {
# Line 1852  for (;;) Line 1857  for (;;)
1857      case OP_NOTI:      case OP_NOTI:
1858      branchlength++;      branchlength++;
1859      cc += 2;      cc += 2;
1860  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32  #ifdef SUPPORT_UTF
1861      if (utf && HAS_EXTRALEN(cc[-1])) cc += GET_EXTRALEN(cc[-1]);      if (utf && HAS_EXTRALEN(cc[-1])) cc += GET_EXTRALEN(cc[-1]);
1862  #endif  #endif
1863      break;      break;
# Line 1864  for (;;) Line 1869  for (;;)
1869      case OP_EXACTI:      case OP_EXACTI:
1870      case OP_NOTEXACT:      case OP_NOTEXACT:
1871      case OP_NOTEXACTI:      case OP_NOTEXACTI:
1872      branchlength += GET2(cc,1);      branchlength += (int)GET2(cc,1);
1873      cc += 2 + IMM2_SIZE;      cc += 2 + IMM2_SIZE;
1874  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32  #ifdef SUPPORT_UTF
1875      if (utf && HAS_EXTRALEN(cc[-1])) cc += GET_EXTRALEN(cc[-1]);      if (utf && HAS_EXTRALEN(cc[-1])) cc += GET_EXTRALEN(cc[-1]);
1876  #endif  #endif
1877      break;      break;
# Line 1932  for (;;) Line 1937  for (;;)
1937        case OP_CRRANGE:        case OP_CRRANGE:
1938        case OP_CRMINRANGE:        case OP_CRMINRANGE:
1939        if (GET2(cc,1) != GET2(cc,1+IMM2_SIZE)) return -1;        if (GET2(cc,1) != GET2(cc,1+IMM2_SIZE)) return -1;
1940        branchlength += GET2(cc,1);        branchlength += (int)GET2(cc,1);
1941        cc += 1 + 2 * IMM2_SIZE;        cc += 1 + 2 * IMM2_SIZE;
1942        break;        break;
1943    
# Line 2059  PRIV(find_bracket)(const pcre_uchar *cod Line 2064  PRIV(find_bracket)(const pcre_uchar *cod
2064  {  {
2065  for (;;)  for (;;)
2066    {    {
2067    register int c = *code;    register pcre_uchar c = *code;
2068    
2069    if (c == OP_END) return NULL;    if (c == OP_END) return NULL;
2070    
# Line 2082  for (;;) Line 2087  for (;;)
2087    else if (c == OP_CBRA || c == OP_SCBRA ||    else if (c == OP_CBRA || c == OP_SCBRA ||
2088             c == OP_CBRAPOS || c == OP_SCBRAPOS)             c == OP_CBRAPOS || c == OP_SCBRAPOS)
2089      {      {
2090      int n = GET2(code, 1+LINK_SIZE);      int n = (int)GET2(code, 1+LINK_SIZE);
2091      if (n == number) return (pcre_uchar *)code;      if (n == number) return (pcre_uchar *)code;
2092      code += PRIV(OP_lengths)[c];      code += PRIV(OP_lengths)[c];
2093      }      }
# Line 2197  find_recurse(const pcre_uchar *code, BOO Line 2202  find_recurse(const pcre_uchar *code, BOO
2202  {  {
2203  for (;;)  for (;;)
2204    {    {
2205    register int c = *code;    register pcre_uchar c = *code;
2206    if (c == OP_END) return NULL;    if (c == OP_END) return NULL;
2207    if (c == OP_RECURSE) return code;    if (c == OP_RECURSE) return code;
2208    
# Line 2351  static BOOL Line 2356  static BOOL
2356  could_be_empty_branch(const pcre_uchar *code, const pcre_uchar *endcode,  could_be_empty_branch(const pcre_uchar *code, const pcre_uchar *endcode,
2357    BOOL utf, compile_data *cd)    BOOL utf, compile_data *cd)
2358  {  {
2359  register int c;  register pcre_uchar c;
2360  for (code = first_significant_code(code + PRIV(OP_lengths)[*code], TRUE);  for (code = first_significant_code(code + PRIV(OP_lengths)[*code], TRUE);
2361       code < endcode;       code < endcode;
2362       code = first_significant_code(code + PRIV(OP_lengths)[c], TRUE))       code = first_significant_code(code + PRIV(OP_lengths)[c], TRUE))
# Line 2385  for (code = first_significant_code(code Line 2390  for (code = first_significant_code(code
2390      /* Test for forward reference */      /* Test for forward reference */
2391    
2392      for (scode = cd->start_workspace; scode < cd->hwm; scode += LINK_SIZE)      for (scode = cd->start_workspace; scode < cd->hwm; scode += LINK_SIZE)
2393        if (GET(scode, 0) == code + 1 - cd->start_code) return TRUE;        if ((int)GET(scode, 0) == (int)(code + 1 - cd->start_code)) return TRUE;
2394    
2395      /* Not a forward reference, test for completed backward reference */      /* Not a forward reference, test for completed backward reference */
2396    
# Line 2706  Returns:   TRUE or FALSE Line 2711  Returns:   TRUE or FALSE
2711  static BOOL  static BOOL
2712  check_posix_syntax(const pcre_uchar *ptr, const pcre_uchar **endptr)  check_posix_syntax(const pcre_uchar *ptr, const pcre_uchar **endptr)
2713  {  {
2714  int terminator;          /* Don't combine these lines; the Solaris cc */  pcre_uchar terminator;          /* Don't combine these lines; the Solaris cc */
2715  terminator = *(++ptr);   /* compiler warns about "non-constant" initializer. */  terminator = *(++ptr);   /* compiler warns about "non-constant" initializer. */
2716  for (++ptr; *ptr != 0; ptr++)  for (++ptr; *ptr != 0; ptr++)
2717    {    {
# Line 2755  register int yield = 0; Line 2760  register int yield = 0;
2760  while (posix_name_lengths[yield] != 0)  while (posix_name_lengths[yield] != 0)
2761    {    {
2762    if (len == posix_name_lengths[yield] &&    if (len == posix_name_lengths[yield] &&
2763      STRNCMP_UC_C8(ptr, pn, len) == 0) return yield;      STRNCMP_UC_C8(ptr, pn, (unsigned int)len) == 0) return yield;
2764    pn += posix_name_lengths[yield] + 1;    pn += posix_name_lengths[yield] + 1;
2765    yield++;    yield++;
2766    }    }
# Line 2810  while ((ptr = (pcre_uchar *)find_recurse Line 2815  while ((ptr = (pcre_uchar *)find_recurse
2815    
2816    for (hc = save_hwm; hc < cd->hwm; hc += LINK_SIZE)    for (hc = save_hwm; hc < cd->hwm; hc += LINK_SIZE)
2817      {      {
2818      offset = GET(hc, 0);      offset = (int)GET(hc, 0);
2819      if (cd->start_code + offset == ptr + 1)      if (cd->start_code + offset == ptr + 1)
2820        {        {
2821        PUT(hc, 0, offset + adjust);        PUT(hc, 0, offset + adjust);
# Line 2823  while ((ptr = (pcre_uchar *)find_recurse Line 2828  while ((ptr = (pcre_uchar *)find_recurse
2828    
2829    if (hc >= cd->hwm)    if (hc >= cd->hwm)
2830      {      {
2831      offset = GET(ptr, 1);      offset = (int)GET(ptr, 1);
2832      if (cd->start_code + offset >= group) PUT(ptr, 1, offset + adjust);      if (cd->start_code + offset >= group) PUT(ptr, 1, offset + adjust);
2833      }      }
2834    
# Line 2913  get_othercase_range(pcre_uint32 *cptr, p Line 2918  get_othercase_range(pcre_uint32 *cptr, p
2918    pcre_uint32 *odptr)    pcre_uint32 *odptr)
2919  {  {
2920  pcre_uint32 c, othercase, next;  pcre_uint32 c, othercase, next;
2921  int co;  unsigned int co;
2922    
2923  /* Find the first character that has an other case. If it has multiple other  /* Find the first character that has an other case. If it has multiple other
2924  cases, return its case offset value. */  cases, return its case offset value. */
# Line 2924  for (c = *cptr; c <= d; c++) Line 2929  for (c = *cptr; c <= d; c++)
2929      {      {
2930      *ocptr = c++;   /* Character that has the set */      *ocptr = c++;   /* Character that has the set */
2931      *cptr = c;      /* Rest of input range */      *cptr = c;      /* Rest of input range */
2932      return co;      return (int)co;
2933      }      }
2934    if ((othercase = UCD_OTHERCASE(c)) != c) break;    if ((othercase = UCD_OTHERCASE(c)) != c) break;
2935    }    }
# Line 2964  Returns:       TRUE if auto-possessifyin Line 2969  Returns:       TRUE if auto-possessifyin
2969  */  */
2970    
2971  static BOOL  static BOOL
2972  check_char_prop(pcre_uint32 c, int ptype, int pdata, BOOL negated)  check_char_prop(pcre_uint32 c, unsigned int ptype, unsigned int pdata, BOOL negated)
2973  {  {
2974  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2975  const pcre_uint32 *p;  const pcre_uint32 *p;
# Line 3053  check_auto_possessive(const pcre_uchar * Line 3058  check_auto_possessive(const pcre_uchar *
3058  pcre_uint32 c = NOTACHAR;  pcre_uint32 c = NOTACHAR;
3059  pcre_uint32 next;  pcre_uint32 next;
3060  int escape;  int escape;
3061  int op_code = *previous++;  pcre_uchar op_code = *previous++;
3062    
3063  /* Skip whitespace and comments in extended mode */  /* Skip whitespace and comments in extended mode */
3064    
# Line 3131  if (*ptr == CHAR_ASTERISK || *ptr == CHA Line 3136  if (*ptr == CHAR_ASTERISK || *ptr == CHA
3136    
3137  if (op_code == OP_CHAR || op_code == OP_CHARI ||  if (op_code == OP_CHAR || op_code == OP_CHARI ||
3138      op_code == OP_NOT || op_code == OP_NOTI)      op_code == OP_NOT || op_code == OP_NOTI)
   //if (escape == 0) switch(op_code)  
3139    {    {
3140  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3141    GETCHARTEST(c, previous);    GETCHARTEST(c, previous);
# Line 3151  if (escape == 0) Line 3155  if (escape == 0)
3155  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3156    if (utf && c != NOTACHAR && (options & PCRE_CASELESS) != 0)    if (utf && c != NOTACHAR && (options & PCRE_CASELESS) != 0)
3157      {      {
3158      int ocs = UCD_CASESET(next);      unsigned int ocs = UCD_CASESET(next);
3159      if (ocs > 0) return check_char_prop(c, PT_CLIST, ocs, op_code >= OP_NOT);      if (ocs > 0) return check_char_prop(c, PT_CLIST, ocs, op_code >= OP_NOT);
3160      }      }
3161  #endif  #endif
# Line 3337  switch(op_code) Line 3341  switch(op_code)
3341      case ESC_p:      case ESC_p:
3342      case ESC_P:      case ESC_P:
3343        {        {
3344        int ptype, pdata, errorcodeptr;        unsigned int ptype = 0, pdata = 0;
3345          int errorcodeptr;
3346        BOOL negated;        BOOL negated;
3347    
3348        ptr--;      /* Make ptr point at the p or P */        ptr--;      /* Make ptr point at the p or P */
3349        ptype = get_ucp(&ptr, &negated, &pdata, &errorcodeptr);        if (!get_ucp(&ptr, &negated, &ptype, &pdata, &errorcodeptr))
3350        if (ptype < 0) return FALSE;          return FALSE;
3351        ptr++;      /* Point past the final curly ket */        ptr++;      /* Point past the final curly ket */
3352    
3353        /* If the property item is optional, we have to give up. (When generated        /* If the property item is optional, we have to give up. (When generated
# Line 3666  Arguments: Line 3671  Arguments:
3671    codeptr        points to the pointer to the current code point    codeptr        points to the pointer to the current code point
3672    ptrptr         points to the current pattern pointer    ptrptr         points to the current pattern pointer
3673    errorcodeptr   points to error code variable    errorcodeptr   points to error code variable
3674    firstcharptr   set to initial literal character, or < 0 (REQ_UNSET, REQ_NONE)    firstcharptr    place to put the first required character
3675    reqcharptr     set to the last literal character required, else < 0    firstcharflagsptr place to put the first character flags, or a negative number
3676      reqcharptr     place to put the last required character
3677      reqcharflagsptr place to put the last required character flags, or a negative number
3678    bcptr          points to current branch chain    bcptr          points to current branch chain
3679    cond_depth     conditional nesting depth    cond_depth     conditional nesting depth
3680    cd             contains pointers to tables etc.    cd             contains pointers to tables etc.
# Line 3680  Returns:         TRUE on success Line 3687  Returns:         TRUE on success
3687    
3688  static BOOL  static BOOL
3689  compile_branch(int *optionsptr, pcre_uchar **codeptr,  compile_branch(int *optionsptr, pcre_uchar **codeptr,
3690    const pcre_uchar **ptrptr, int *errorcodeptr, pcre_int32 *firstcharptr,    const pcre_uchar **ptrptr, int *errorcodeptr,
3691    pcre_int32 *reqcharptr, branch_chain *bcptr, int cond_depth,    pcre_uint32 *firstcharptr, pcre_int32 *firstcharflagsptr,
3692      pcre_uint32 *reqcharptr, pcre_int32 *reqcharflagsptr,
3693      branch_chain *bcptr, int cond_depth,
3694    compile_data *cd, int *lengthptr)    compile_data *cd, int *lengthptr)
3695  {  {
3696  int repeat_type, op_type;  int repeat_type, op_type;
3697  int repeat_min = 0, repeat_max = 0;      /* To please picky compilers */  int repeat_min = 0, repeat_max = 0;      /* To please picky compilers */
3698  int bravalue = 0;  int bravalue = 0;
3699  int greedy_default, greedy_non_default;  int greedy_default, greedy_non_default;
3700  pcre_int32 firstchar, reqchar;  pcre_uint32 firstchar, reqchar;
3701  pcre_int32 zeroreqchar, zerofirstchar;  pcre_int32 firstcharflags, reqcharflags;
3702    pcre_uint32 zeroreqchar, zerofirstchar;
3703    pcre_int32 zeroreqcharflags, zerofirstcharflags;
3704  pcre_int32 req_caseopt, reqvary, tempreqvary;  pcre_int32 req_caseopt, reqvary, tempreqvary;
3705  int options = *optionsptr;               /* May change dynamically */  int options = *optionsptr;               /* May change dynamically */
3706  int after_manual_callout = 0;  int after_manual_callout = 0;
# Line 3717  dynamically as we process the pattern. * Line 3728  dynamically as we process the pattern. *
3728  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3729  /* PCRE_UTF[16|32] have the same value as PCRE_UTF8. */  /* PCRE_UTF[16|32] have the same value as PCRE_UTF8. */
3730  BOOL utf = (options & PCRE_UTF8) != 0;  BOOL utf = (options & PCRE_UTF8) != 0;
3731    #ifndef COMPILE_PCRE32
3732  pcre_uchar utf_chars[6];  pcre_uchar utf_chars[6];
3733    #endif
3734  #else  #else
3735  BOOL utf = FALSE;  BOOL utf = FALSE;
3736  #endif  #endif
# Line 3752  to take the zero repeat into account. Th Line 3765  to take the zero repeat into account. Th
3765  zerofirstbyte and zeroreqchar when such a repeat is encountered. The individual  zerofirstbyte and zeroreqchar when such a repeat is encountered. The individual
3766  item types that can be repeated set these backoff variables appropriately. */  item types that can be repeated set these backoff variables appropriately. */
3767    
3768  firstchar = reqchar = zerofirstchar = zeroreqchar = REQ_UNSET;  firstchar = reqchar = zerofirstchar = zeroreqchar = 0;
3769    firstcharflags = reqcharflags = zerofirstcharflags = zeroreqcharflags = REQ_UNSET;
3770    
3771  /* The variable req_caseopt contains either the REQ_CASELESS value  /* The variable req_caseopt contains either the REQ_CASELESS value
3772  or zero, according to the current setting of the caseless flag. The  or zero, according to the current setting of the caseless flag. The
# Line 3778  for (;; ptr++) Line 3792  for (;; ptr++)
3792    int recno;    int recno;
3793    int refsign;    int refsign;
3794    int skipbytes;    int skipbytes;
3795    int subreqchar;    pcre_uint32 subreqchar, subfirstchar;
3796    int subfirstchar;    pcre_int32 subreqcharflags, subfirstcharflags;
3797    int terminator;    int terminator;
3798    int mclength;    unsigned int mclength;
3799    int tempbracount;    int tempbracount;
3800    int ec; // FIXMEchpe pcre_uint32    pcre_uint32 ec;
3801    pcre_uchar mcbuffer[8];    pcre_uchar mcbuffer[8];
3802    
3803    /* Get next character in the pattern */    /* Get next character in the pattern */
# Line 3946  for (;; ptr++) Line 3960  for (;; ptr++)
3960      case CHAR_VERTICAL_LINE:       /* or | or ) */      case CHAR_VERTICAL_LINE:       /* or | or ) */
3961      case CHAR_RIGHT_PARENTHESIS:      case CHAR_RIGHT_PARENTHESIS:
3962      *firstcharptr = firstchar;      *firstcharptr = firstchar;
3963        *firstcharflagsptr = firstcharflags;
3964      *reqcharptr = reqchar;      *reqcharptr = reqchar;
3965        *reqcharflagsptr = reqcharflags;
3966      *codeptr = code;      *codeptr = code;
3967      *ptrptr = ptr;      *ptrptr = ptr;
3968      if (lengthptr != NULL)      if (lengthptr != NULL)
# Line 3970  for (;; ptr++) Line 3986  for (;; ptr++)
3986      previous = NULL;      previous = NULL;
3987      if ((options & PCRE_MULTILINE) != 0)      if ((options & PCRE_MULTILINE) != 0)
3988        {        {
3989        if (firstchar == REQ_UNSET) firstchar = REQ_NONE;        if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
3990        *code++ = OP_CIRCM;        *code++ = OP_CIRCM;
3991        }        }
3992      else *code++ = OP_CIRC;      else *code++ = OP_CIRC;
# Line 3985  for (;; ptr++) Line 4001  for (;; ptr++)
4001      repeats. The value of reqchar doesn't change either. */      repeats. The value of reqchar doesn't change either. */
4002    
4003      case CHAR_DOT:      case CHAR_DOT:
4004      if (firstchar == REQ_UNSET) firstchar = REQ_NONE;      if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
4005      zerofirstchar = firstchar;      zerofirstchar = firstchar;
4006        zerofirstcharflags = firstcharflags;
4007      zeroreqchar = reqchar;      zeroreqchar = reqchar;
4008        zeroreqcharflags = reqcharflags;
4009      previous = code;      previous = code;
4010      *code++ = ((options & PCRE_DOTALL) != 0)? OP_ALLANY: OP_ANY;      *code++ = ((options & PCRE_DOTALL) != 0)? OP_ALLANY: OP_ANY;
4011      break;      break;
# Line 4061  for (;; ptr++) Line 4079  for (;; ptr++)
4079          (cd->external_options & PCRE_JAVASCRIPT_COMPAT) != 0)          (cd->external_options & PCRE_JAVASCRIPT_COMPAT) != 0)
4080        {        {
4081        *code++ = negate_class? OP_ALLANY : OP_FAIL;        *code++ = negate_class? OP_ALLANY : OP_FAIL;
4082        if (firstchar == REQ_UNSET) firstchar = REQ_NONE;        if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
4083        zerofirstchar = firstchar;        zerofirstchar = firstchar;
4084          zerofirstcharflags = firstcharflags;
4085        break;        break;
4086        }        }
4087    
# Line 4101  for (;; ptr++) Line 4120  for (;; ptr++)
4120        {        {
4121        const pcre_uchar *oldptr;        const pcre_uchar *oldptr;
4122    
4123  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32  #ifdef SUPPORT_UTF
4124        if (utf && HAS_EXTRALEN(c))        if (utf && HAS_EXTRALEN(c))
4125          {                           /* Braces are required because the */          {                           /* Braces are required because the */
4126          GETCHARLEN(c, ptr, ptr);    /* macro generates multiple statements */          GETCHARLEN(c, ptr, ptr);    /* macro generates multiple statements */
# Line 4360  for (;; ptr++) Line 4379  for (;; ptr++)
4379              case ESC_P:              case ESC_P:
4380                {                {
4381                BOOL negated;                BOOL negated;
4382                int pdata;                unsigned int ptype = 0, pdata = 0;
4383                int ptype = get_ucp(&ptr, &negated, &pdata, errorcodeptr);                if (!get_ucp(&ptr, &negated, &ptype, &pdata, errorcodeptr))
4384                if (ptype < 0) goto FAILED;                  goto FAILED;
4385                *class_uchardata++ = ((escape == ESC_p) != negated)?                *class_uchardata++ = ((escape == ESC_p) != negated)?
4386                  XCL_PROP : XCL_NOTPROP;                  XCL_PROP : XCL_NOTPROP;
4387                *class_uchardata++ = ptype;                *class_uchardata++ = ptype;
# Line 4416  for (;; ptr++) Line 4435  for (;; ptr++)
4435    
4436        if (!inescq && ptr[1] == CHAR_MINUS)        if (!inescq && ptr[1] == CHAR_MINUS)
4437          {          {
4438          int d;          pcre_uint32 d;
4439          ptr += 2;          ptr += 2;
4440          while (*ptr == CHAR_BACKSLASH && ptr[1] == CHAR_E) ptr += 2;          while (*ptr == CHAR_BACKSLASH && ptr[1] == CHAR_E) ptr += 2;
4441    
# Line 4522  for (;; ptr++) Line 4541  for (;; ptr++)
4541          {          {
4542          ptr++;          ptr++;
4543          zeroreqchar = reqchar;          zeroreqchar = reqchar;
4544            zeroreqcharflags = reqcharflags;
4545    
4546          if (negate_class)          if (negate_class)
4547            {            {
4548  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
             // FIXMEchpe pcreuint32?  
4549            int d;            int d;
4550  #endif  #endif
4551            if (firstchar == REQ_UNSET) firstchar = REQ_NONE;            if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
4552            zerofirstchar = firstchar;            zerofirstchar = firstchar;
4553              zerofirstcharflags = firstcharflags;
4554    
4555            /* For caseless UTF-8 mode when UCP support is available, check            /* For caseless UTF-8 mode when UCP support is available, check
4556            whether this character has more than one other case. If so, generate            whether this character has more than one other case. If so, generate
# Line 4618  for (;; ptr++) Line 4638  for (;; ptr++)
4638      setting, whatever the repeat count. Any reqchar setting must remain      setting, whatever the repeat count. Any reqchar setting must remain
4639      unchanged after any kind of repeat. */      unchanged after any kind of repeat. */
4640    
4641      if (firstchar == REQ_UNSET) firstchar = REQ_NONE;      if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
4642      zerofirstchar = firstchar;      zerofirstchar = firstchar;
4643        zerofirstcharflags = firstcharflags;
4644      zeroreqchar = reqchar;      zeroreqchar = reqchar;
4645        zeroreqcharflags = reqcharflags;
4646    
4647      /* If there are characters with values > 255, we have to compile an      /* If there are characters with values > 255, we have to compile an
4648      extended class, with its own opcode, unless there was a negated special      extended class, with its own opcode, unless there was a negated special
# Line 4715  for (;; ptr++) Line 4737  for (;; ptr++)
4737      if (repeat_min == 0)      if (repeat_min == 0)
4738        {        {
4739        firstchar = zerofirstchar;    /* Adjust for zero repeat */        firstchar = zerofirstchar;    /* Adjust for zero repeat */
4740          firstcharflags = zerofirstcharflags;
4741        reqchar = zeroreqchar;        /* Ditto */        reqchar = zeroreqchar;        /* Ditto */
4742          reqcharflags = zeroreqcharflags;
4743        }        }
4744    
4745      /* Remember whether this is a variable length repeat */      /* Remember whether this is a variable length repeat */
# Line 4818  for (;; ptr++) Line 4842  for (;; ptr++)
4842          {          {
4843          c = code[-1];          c = code[-1];
4844          if (*previous <= OP_CHARI && repeat_min > 1)          if (*previous <= OP_CHARI && repeat_min > 1)
4845            reqchar = c | req_caseopt | cd->req_varyopt;            {
4846              reqchar = c;
4847              reqcharflags = req_caseopt | cd->req_varyopt;
4848              }
4849          }          }
4850    
4851        /* If the repetition is unlimited, it pays to see if the next thing on        /* If the repetition is unlimited, it pays to see if the next thing on
# Line 5200  for (;; ptr++) Line 5227  for (;; ptr++)
5227    
5228            else            else
5229              {              {
5230              if (groupsetfirstchar && reqchar < 0) reqchar = firstchar;              if (groupsetfirstchar && reqcharflags < 0)
5231                  {
5232                  reqchar = firstchar;
5233                  reqcharflags = firstcharflags;
5234                  }
5235    
5236              for (i = 1; i < repeat_min; i++)              for (i = 1; i < repeat_min; i++)
5237                {                {
# Line 5479  for (;; ptr++) Line 5510  for (;; ptr++)
5510        else if (*tempcode == OP_EXACT || *tempcode == OP_NOTEXACT)        else if (*tempcode == OP_EXACT || *tempcode == OP_NOTEXACT)
5511          {          {
5512          tempcode += PRIV(OP_lengths)[*tempcode];          tempcode += PRIV(OP_lengths)[*tempcode];
5513  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32  #ifdef SUPPORT_UTF
5514          if (utf && HAS_EXTRALEN(tempcode[-1]))          if (utf && HAS_EXTRALEN(tempcode[-1]))
5515            tempcode += GET_EXTRALEN(tempcode[-1]);            tempcode += GET_EXTRALEN(tempcode[-1]);
5516  #endif  #endif
# Line 5620  for (;; ptr++) Line 5651  for (;; ptr++)
5651                (cd->assert_depth > 0)? OP_ASSERT_ACCEPT : OP_ACCEPT;                (cd->assert_depth > 0)? OP_ASSERT_ACCEPT : OP_ACCEPT;
5652    
5653              /* Do not set firstchar after *ACCEPT */              /* Do not set firstchar after *ACCEPT */
5654              if (firstchar == REQ_UNSET) firstchar = REQ_NONE;              if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
5655              }              }
5656    
5657            /* Handle other cases with/without an argument */            /* Handle other cases with/without an argument */
# Line 5786  for (;; ptr++) Line 5817  for (;; ptr++)
5817          while (MAX_255(*ptr) && (cd->ctypes[*ptr] & ctype_word) != 0)          while (MAX_255(*ptr) && (cd->ctypes[*ptr] & ctype_word) != 0)
5818            {            {
5819            if (recno >= 0)            if (recno >= 0)
5820              recno = (IS_DIGIT(*ptr))? recno * 10 + *ptr - CHAR_0 : -1;              recno = (IS_DIGIT(*ptr))? recno * 10 + (int)(*ptr - CHAR_0) : -1;
5821            ptr++;            ptr++;
5822            }            }
5823          namelen = (int)(ptr - name);          namelen = (int)(ptr - name);
5824    
5825          if ((terminator > 0 && *ptr++ != terminator) ||          if ((terminator > 0 && *ptr++ != (pcre_uchar)terminator) ||
5826              *ptr++ != CHAR_RIGHT_PARENTHESIS)              *ptr++ != CHAR_RIGHT_PARENTHESIS)
5827            {            {
5828            ptr--;      /* Error offset */            ptr--;      /* Error offset */
# Line 6031  for (;; ptr++) Line 6062  for (;; ptr++)
6062    
6063            if (lengthptr != NULL)            if (lengthptr != NULL)
6064              {              {
6065              if (*ptr != terminator)              if (*ptr != (pcre_uchar)terminator)
6066                {                {
6067                *errorcodeptr = ERR42;                *errorcodeptr = ERR42;
6068                goto FAILED;                goto FAILED;
# Line 6173  for (;; ptr++) Line 6204  for (;; ptr++)
6204              *errorcodeptr = ERR62;              *errorcodeptr = ERR62;
6205              goto FAILED;              goto FAILED;
6206              }              }
6207            if (*ptr != terminator)            if (*ptr != (pcre_uchar)terminator)
6208              {              {
6209              *errorcodeptr = ERR42;              *errorcodeptr = ERR42;
6210              goto FAILED;              goto FAILED;
# Line 6279  for (;; ptr++) Line 6310  for (;; ptr++)
6310            while(IS_DIGIT(*ptr))            while(IS_DIGIT(*ptr))
6311              recno = recno * 10 + *ptr++ - CHAR_0;              recno = recno * 10 + *ptr++ - CHAR_0;
6312    
6313            if (*ptr != terminator)            if (*ptr != (pcre_uchar)terminator)
6314              {              {
6315              *errorcodeptr = ERR29;              *errorcodeptr = ERR29;
6316              goto FAILED;              goto FAILED;
# Line 6383  for (;; ptr++) Line 6414  for (;; ptr++)
6414    
6415          /* Can't determine a first byte now */          /* Can't determine a first byte now */
6416    
6417          if (firstchar == REQ_UNSET) firstchar = REQ_NONE;          if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
6418          continue;          continue;
6419    
6420    
# Line 6517  for (;; ptr++) Line 6548  for (;; ptr++)
6548           cond_depth +           cond_depth +
6549             ((bravalue == OP_COND)?1:0),   /* Depth of condition subpatterns */             ((bravalue == OP_COND)?1:0),   /* Depth of condition subpatterns */
6550           &subfirstchar,                   /* For possible first char */           &subfirstchar,                   /* For possible first char */
6551             &subfirstcharflags,
6552           &subreqchar,                     /* For possible last char */           &subreqchar,                     /* For possible last char */
6553             &subreqcharflags,
6554           bcptr,                           /* Current branch chain */           bcptr,                           /* Current branch chain */
6555           cd,                              /* Tables block */           cd,                              /* Tables block */
6556           (lengthptr == NULL)? NULL :      /* Actual compile phase */           (lengthptr == NULL)? NULL :      /* Actual compile phase */
# Line 6578  for (;; ptr++) Line 6611  for (;; ptr++)
6611            *errorcodeptr = ERR27;            *errorcodeptr = ERR27;
6612            goto FAILED;            goto FAILED;
6613            }            }
6614          if (condcount == 1) subfirstchar = subreqchar = REQ_NONE;          if (condcount == 1) subfirstcharflags = subreqcharflags = REQ_NONE;
6615          }          }
6616        }        }
6617    
# Line 6627  for (;; ptr++) Line 6660  for (;; ptr++)
6660      back off. */      back off. */
6661    
6662      zeroreqchar = reqchar;      zeroreqchar = reqchar;
6663        zeroreqcharflags = reqcharflags;
6664      zerofirstchar = firstchar;      zerofirstchar = firstchar;
6665        zerofirstcharflags = firstcharflags;
6666      groupsetfirstchar = FALSE;      groupsetfirstchar = FALSE;
6667    
6668      if (bravalue >= OP_ONCE)      if (bravalue >= OP_ONCE)
# Line 6638  for (;; ptr++) Line 6673  for (;; ptr++)
6673        no firstchar, set "none" for the whole branch. In both cases, a zero        no firstchar, set "none" for the whole branch. In both cases, a zero
6674        repeat forces firstchar to "none". */        repeat forces firstchar to "none". */
6675    
6676        if (firstchar == REQ_UNSET)        if (firstcharflags == REQ_UNSET)
6677          {          {
6678          if (subfirstchar >= 0)          if (subfirstcharflags >= 0)
6679            {            {
6680            firstchar = subfirstchar;            firstchar = subfirstchar;
6681              firstcharflags = subfirstcharflags;
6682            groupsetfirstchar = TRUE;            groupsetfirstchar = TRUE;
6683            }            }
6684          else firstchar = REQ_NONE;          else firstcharflags = REQ_NONE;
6685          zerofirstchar = REQ_NONE;          zerofirstcharflags = REQ_NONE;
6686          }          }
6687    
6688        /* If firstchar was previously set, convert the subpattern's firstchar        /* If firstchar was previously set, convert the subpattern's firstchar
6689        into reqchar if there wasn't one, using the vary flag that was in        into reqchar if there wasn't one, using the vary flag that was in
6690        existence beforehand. */        existence beforehand. */
6691    
6692        else if (subfirstchar >= 0 && subreqchar < 0)        else if (subfirstcharflags >= 0 && subreqcharflags < 0)
6693          subreqchar = subfirstchar | tempreqvary;          {
6694            subreqchar = subfirstchar;
6695            subreqcharflags = subfirstcharflags | tempreqvary;
6696            }
6697    
6698        /* If the subpattern set a required byte (or set a first byte that isn't        /* If the subpattern set a required byte (or set a first byte that isn't
6699        really the first byte - see above), set it. */        really the first byte - see above), set it. */
6700    
6701        if (subreqchar >= 0) reqchar = subreqchar;        if (subreqcharflags >= 0)
6702            {
6703            reqchar = subreqchar;
6704            reqcharflags = subreqcharflags;
6705            }
6706        }        }
6707    
6708      /* For a forward assertion, we take the reqchar, if set. This can be      /* For a forward assertion, we take the reqchar, if set. This can be
# Line 6670  for (;; ptr++) Line 6713  for (;; ptr++)
6713      of a firstchar. This is overcome by a scan at the end if there's no      of a firstchar. This is overcome by a scan at the end if there's no
6714      firstchar, looking for an asserted first char. */      firstchar, looking for an asserted first char. */
6715    
6716      else if (bravalue == OP_ASSERT && subreqchar >= 0) reqchar = subreqchar;      else if (bravalue == OP_ASSERT && subreqcharflags >= 0)
6717          {
6718          reqchar = subreqchar;
6719          reqcharflags = subreqcharflags;
6720          }
6721      break;     /* End of processing '(' */      break;     /* End of processing '(' */
6722    
6723    
# Line 6706  for (;; ptr++) Line 6753  for (;; ptr++)
6753        /* For metasequences that actually match a character, we disable the        /* For metasequences that actually match a character, we disable the
6754        setting of a first character if it hasn't already been set. */        setting of a first character if it hasn't already been set. */
6755    
6756        if (firstchar == REQ_UNSET && escape > ESC_b && escape < ESC_Z)        if (firstcharflags == REQ_UNSET && escape > ESC_b && escape < ESC_Z)
6757          firstchar = REQ_NONE;          firstcharflags = REQ_NONE;
6758    
6759        /* Set values to reset to if this is followed by a zero repeat. */        /* Set values to reset to if this is followed by a zero repeat. */
6760    
6761        zerofirstchar = firstchar;        zerofirstchar = firstchar;
6762          zerofirstcharflags = firstcharflags;
6763        zeroreqchar = reqchar;        zeroreqchar = reqchar;
6764          zeroreqcharflags = reqcharflags;
6765    
6766        /* \g<name> or \g'name' is a subroutine call by name and \g<n> or \g'n'        /* \g<name> or \g'name' is a subroutine call by name and \g<n> or \g'n'
6767        is a subroutine call by number (Oniguruma syntax). In fact, the value        is a subroutine call by number (Oniguruma syntax). In fact, the value
# Line 6741  for (;; ptr++) Line 6790  for (;; ptr++)
6790          if (ptr[1] != CHAR_PLUS && ptr[1] != CHAR_MINUS)          if (ptr[1] != CHAR_PLUS && ptr[1] != CHAR_MINUS)
6791            {            {
6792            BOOL is_a_number = TRUE;            BOOL is_a_number = TRUE;
6793            for (p = ptr + 1; *p != 0 && *p != terminator; p++)            for (p = ptr + 1; *p != 0 && *p != (pcre_uchar)terminator; p++)
6794              {              {
6795              if (!MAX_255(*p)) { is_a_number = FALSE; break; }              if (!MAX_255(*p)) { is_a_number = FALSE; break; }
6796              if ((cd->ctypes[*p] & ctype_digit) == 0) is_a_number = FALSE;              if ((cd->ctypes[*p] & ctype_digit) == 0) is_a_number = FALSE;
6797              if ((cd->ctypes[*p] & ctype_word) == 0) break;              if ((cd->ctypes[*p] & ctype_word) == 0) break;
6798              }              }
6799            if (*p != terminator)            if (*p != (pcre_uchar)terminator)
6800              {              {
6801              *errorcodeptr = ERR57;              *errorcodeptr = ERR57;
6802              break;              break;
# Line 6765  for (;; ptr++) Line 6814  for (;; ptr++)
6814    
6815          p = ptr + 2;          p = ptr + 2;
6816          while (IS_DIGIT(*p)) p++;          while (IS_DIGIT(*p)) p++;
6817          if (*p != terminator)          if (*p != (pcre_uchar)terminator)
6818            {            {
6819            *errorcodeptr = ERR57;            *errorcodeptr = ERR57;
6820            break;            break;
# Line 6802  for (;; ptr++) Line 6851  for (;; ptr++)
6851          recno = -escape;          recno = -escape;
6852    
6853          HANDLE_REFERENCE:    /* Come here from named backref handling */          HANDLE_REFERENCE:    /* Come here from named backref handling */
6854          if (firstchar == REQ_UNSET) firstchar = REQ_NONE;          if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
6855          previous = code;          previous = code;
6856          *code++ = ((options & PCRE_CASELESS) != 0)? OP_REFI : OP_REF;          *code++ = ((options & PCRE_CASELESS) != 0)? OP_REFI : OP_REF;
6857          PUT2INC(code, 0, recno);          PUT2INC(code, 0, recno);
# Line 6829  for (;; ptr++) Line 6878  for (;; ptr++)
6878        else if (escape == ESC_P || escape == ESC_p)        else if (escape == ESC_P || escape == ESC_p)
6879          {          {
6880          BOOL negated;          BOOL negated;
6881          int pdata;          unsigned int ptype = 0, pdata = 0;
6882          int ptype = get_ucp(&ptr, &negated, &pdata, errorcodeptr);          if (!get_ucp(&ptr, &negated, &ptype, &pdata, errorcodeptr))
6883          if (ptype < 0) goto FAILED;            goto FAILED;
6884          previous = code;          previous = code;
6885          *code++ = ((escape == ESC_p) != negated)? OP_PROP : OP_NOTPROP;          *code++ = ((escape == ESC_p) != negated)? OP_PROP : OP_NOTPROP;
6886          *code++ = ptype;          *code++ = ptype;
# Line 6905  for (;; ptr++) Line 6954  for (;; ptr++)
6954      mclength = 1;      mclength = 1;
6955      mcbuffer[0] = c;      mcbuffer[0] = c;
6956    
6957  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32  #ifdef SUPPORT_UTF
6958      if (utf && HAS_EXTRALEN(c))      if (utf && HAS_EXTRALEN(c))
6959        ACROSSCHAR(TRUE, ptr[1], mcbuffer[mclength++] = *(++ptr));        ACROSSCHAR(TRUE, ptr[1], mcbuffer[mclength++] = *(++ptr));
6960  #endif  #endif
# Line 6929  for (;; ptr++) Line 6978  for (;; ptr++)
6978          *code++ = OP_PROP;          *code++ = OP_PROP;
6979          *code++ = PT_CLIST;          *code++ = PT_CLIST;
6980          *code++ = c;          *code++ = c;
6981          if (firstchar == REQ_UNSET) firstchar = zerofirstchar = REQ_NONE;          if (firstcharflags == REQ_UNSET) firstcharflags = zerofirstcharflags = REQ_NONE;
6982          break;          break;
6983          }          }
6984        }        }
# Line 6950  for (;; ptr++) Line 6999  for (;; ptr++)
6999      Otherwise, leave the firstchar value alone, and don't change it on a zero      Otherwise, leave the firstchar value alone, and don't change it on a zero
7000      repeat. */      repeat. */
7001    
7002      if (firstchar == REQ_UNSET)      if (firstcharflags == REQ_UNSET)
7003        {        {
7004        zerofirstchar = REQ_NONE;        zerofirstcharflags = REQ_NONE;
7005        zeroreqchar = reqchar;        zeroreqchar = reqchar;
7006          zeroreqcharflags = reqcharflags;
7007    
7008        /* If the character is more than one byte long, we can set firstchar        /* If the character is more than one byte long, we can set firstchar
7009        only if it is not to be matched caselessly. */        only if it is not to be matched caselessly. */
# Line 6961  for (;; ptr++) Line 7011  for (;; ptr++)
7011        if (mclength == 1 || req_caseopt == 0)        if (mclength == 1 || req_caseopt == 0)
7012          {          {
7013          firstchar = mcbuffer[0] | req_caseopt;          firstchar = mcbuffer[0] | req_caseopt;
7014          if (mclength != 1) reqchar = code[-1] | cd->req_varyopt;          firstchar = mcbuffer[0];
7015            firstcharflags = req_caseopt;
7016    
7017            if (mclength != 1)
7018              {
7019              reqchar = code[-1];
7020              reqcharflags = cd->req_varyopt;
7021              }
7022          }          }
7023        else firstchar = reqchar = REQ_NONE;        else firstcharflags = reqcharflags = REQ_NONE;
7024        }        }
7025    
7026      /* firstchar was previously set; we can set reqchar only if the length is      /* firstchar was previously set; we can set reqchar only if the length is
# Line 6972  for (;; ptr++) Line 7029  for (;; ptr++)
7029      else      else
7030        {        {
7031        zerofirstchar = firstchar;        zerofirstchar = firstchar;
7032          zerofirstcharflags = firstcharflags;
7033        zeroreqchar = reqchar;        zeroreqchar = reqchar;
7034          zeroreqcharflags = reqcharflags;
7035        if (mclength == 1 || req_caseopt == 0)        if (mclength == 1 || req_caseopt == 0)
7036          reqchar = code[-1] | req_caseopt | cd->req_varyopt;          {
7037            reqchar = code[-1];
7038            reqcharflags = req_caseopt | cd->req_varyopt;
7039            }
7040        }        }
7041    
7042      break;            /* End of literal character handling */      break;            /* End of literal character handling */
# Line 6993  return FALSE; Line 7055  return FALSE;
7055    
7056    
7057    
   
7058  /*************************************************  /*************************************************
7059  *     Compile sequence of alternatives           *  *     Compile sequence of alternatives           *
7060  *************************************************/  *************************************************/
# Line 7014  Arguments: Line 7075  Arguments:
7075    reset_bracount TRUE to reset the count for each branch    reset_bracount TRUE to reset the count for each branch
7076    skipbytes      skip this many bytes at start (for brackets and OP_COND)    skipbytes      skip this many bytes at start (for brackets and OP_COND)
7077    cond_depth     depth of nesting for conditional subpatterns    cond_depth     depth of nesting for conditional subpatterns
7078    firstcharptr   place to put the first required character, or a negative number    firstcharptr    place to put the first required character
7079    reqcharptr     place to put the last required character, or a negative number    firstcharflagsptr place to put the first character flags, or a negative number
7080      reqcharptr     place to put the last required character
7081      reqcharflagsptr place to put the last required character flags, or a negative number
7082    bcptr          pointer to the chain of currently open branches    bcptr          pointer to the chain of currently open branches
7083    cd             points to the data block with tables pointers etc.    cd             points to the data block with tables pointers etc.
7084    lengthptr      NULL during the real compile phase    lengthptr      NULL during the real compile phase
# Line 7027  Returns:         TRUE on success Line 7090  Returns:         TRUE on success
7090  static BOOL  static BOOL
7091  compile_regex(int options, pcre_uchar **codeptr, const pcre_uchar **ptrptr,  compile_regex(int options, pcre_uchar **codeptr, const pcre_uchar **ptrptr,
7092    int *errorcodeptr, BOOL lookbehind, BOOL reset_bracount, int skipbytes,    int *errorcodeptr, BOOL lookbehind, BOOL reset_bracount, int skipbytes,
7093    int cond_depth, pcre_int32 *firstcharptr, pcre_int32 *reqcharptr,    int cond_depth,
7094      pcre_uint32 *firstcharptr, pcre_int32 *firstcharflagsptr,
7095      pcre_uint32 *reqcharptr, pcre_int32 *reqcharflagsptr,
7096    branch_chain *bcptr, compile_data *cd, int *lengthptr)    branch_chain *bcptr, compile_data *cd, int *lengthptr)
7097  {  {
7098  const pcre_uchar *ptr = *ptrptr;  const pcre_uchar *ptr = *ptrptr;
# Line 7037  pcre_uchar *start_bracket = code; Line 7102  pcre_uchar *start_bracket = code;
7102  pcre_uchar *reverse_count = NULL;  pcre_uchar *reverse_count = NULL;
7103  open_capitem capitem;  open_capitem capitem;
7104  int capnumber = 0;  int capnumber = 0;
7105  pcre_int32 firstchar, reqchar;  pcre_uint32 firstchar, reqchar;
7106  pcre_int32 branchfirstchar, branchreqchar;  pcre_int32 firstcharflags, reqcharflags;
7107    pcre_uint32 branchfirstchar, branchreqchar;
7108    pcre_int32 branchfirstcharflags, branchreqcharflags;
7109  int length;  int length;
7110  int orig_bracount;  int orig_bracount;
7111  int max_bracount;  int max_bracount;
# Line 7047  branch_chain bc; Line 7114  branch_chain bc;
7114  bc.outer = bcptr;  bc.outer = bcptr;
7115  bc.current_branch = code;  bc.current_branch = code;
7116    
7117  firstchar = reqchar = REQ_UNSET;  firstchar = reqchar = 0;
7118    firstcharflags = reqcharflags = REQ_UNSET;
7119    
7120  /* Accumulate the length for use in the pre-compile phase. Start with the  /* Accumulate the length for use in the pre-compile phase. Start with the
7121  length of the BRA and KET and any extra bytes that are required at the  length of the BRA and KET and any extra bytes that are required at the
# Line 7107  for (;;) Line 7175  for (;;)
7175    into the length. */    into the length. */
7176    
7177    if (!compile_branch(&options, &code, &ptr, errorcodeptr, &branchfirstchar,    if (!compile_branch(&options, &code, &ptr, errorcodeptr, &branchfirstchar,
7178          &branchreqchar, &bc, cond_depth, cd,          &branchfirstcharflags, &branchreqchar, &branchreqcharflags, &bc,
7179          (lengthptr == NULL)? NULL : &length))          cond_depth, cd, (lengthptr == NULL)? NULL : &length))
7180      {      {
7181      *ptrptr = ptr;      *ptrptr = ptr;
7182      return FALSE;      return FALSE;
# Line 7129  for (;;) Line 7197  for (;;)
7197      if (*last_branch != OP_ALT)      if (*last_branch != OP_ALT)
7198        {        {
7199        firstchar = branchfirstchar;        firstchar = branchfirstchar;
7200          firstcharflags = branchfirstcharflags;
7201        reqchar = branchreqchar;        reqchar = branchreqchar;
7202          reqcharflags = branchreqcharflags;
7203        }        }
7204    
7205      /* If this is not the first branch, the first char and reqchar have to      /* If this is not the first branch, the first char and reqchar have to
# Line 7143  for (;;) Line 7213  for (;;)
7213        we have to abandon the firstchar for the regex, but if there was        we have to abandon the firstchar for the regex, but if there was
7214        previously no reqchar, it takes on the value of the old firstchar. */        previously no reqchar, it takes on the value of the old firstchar. */
7215    
7216        if (firstchar >= 0 && firstchar != branchfirstchar)        if (firstcharflags >= 0 &&
7217              (firstcharflags != branchfirstcharflags || firstchar != branchfirstchar))
7218          {          {
7219          if (reqchar < 0) reqchar = firstchar;          if (reqcharflags < 0)
7220          firstchar = REQ_NONE;            {
7221              reqchar = firstchar;
7222              reqcharflags = firstcharflags;
7223              }
7224            firstcharflags = REQ_NONE;
7225          }          }
7226    
7227        /* If we (now or from before) have no firstchar, a firstchar from the        /* If we (now or from before) have no firstchar, a firstchar from the
7228        branch becomes a reqchar if there isn't a branch reqchar. */        branch becomes a reqchar if there isn't a branch reqchar. */
7229    
7230        if (firstchar < 0 && branchfirstchar >= 0 && branchreqchar < 0)        if (firstcharflags < 0 && branchfirstcharflags >= 0 && branchreqcharflags < 0)
7231            branchreqchar = branchfirstchar;          {
7232            branchreqchar = branchfirstchar;
7233            branchreqcharflags = branchfirstcharflags;
7234            }
7235    
7236        /* Now ensure that the reqchars match */        /* Now ensure that the reqchars match */
7237    
7238        if ((reqchar & ~REQ_VARY) != (branchreqchar & ~REQ_VARY))        if (((reqcharflags & ~REQ_VARY) != (branchreqcharflags & ~REQ_VARY)) ||
7239          reqchar = REQ_NONE;            reqchar != branchreqchar)
7240        else reqchar |= branchreqchar;   /* To "or" REQ_VARY */          reqcharflags = REQ_NONE;
7241          else
7242            {
7243            reqchar = branchreqchar;
7244            reqcharflags |= branchreqcharflags; /* To "or" REQ_VARY */
7245            }
7246        }        }
7247    
7248      /* If lookbehind, check that this branch matches a fixed-length string, and      /* If lookbehind, check that this branch matches a fixed-length string, and
# Line 7255  for (;;) Line 7338  for (;;)
7338      *codeptr = code;      *codeptr = code;
7339      *ptrptr = ptr;      *ptrptr = ptr;
7340      *firstcharptr = firstchar;      *firstcharptr = firstchar;
7341        *firstcharflagsptr = firstcharflags;
7342      *reqcharptr = reqchar;      *reqcharptr = reqchar;
7343        *reqcharflagsptr = reqcharflags;
7344      if (lengthptr != NULL)      if (lengthptr != NULL)
7345        {        {
7346        if (OFLOW_MAX - *lengthptr < length)        if (OFLOW_MAX - *lengthptr < length)
# Line 7542  we return that char, otherwise -1. Line 7627  we return that char, otherwise -1.
7627    
7628  Arguments:  Arguments:
7629    code       points to start of expression (the bracket)    code       points to start of expression (the bracket)
7630      flags       points to the first char flags, or to REQ_NONE
7631    inassert   TRUE if in an assertion    inassert   TRUE if in an assertion
7632    
7633  Returns:     -1 or the fixed first char  Returns:     the fixed first char, or 0 with REQ_NONE in flags
7634  */  */
7635    
7636  static int  static pcre_uint32
7637  find_firstassertedchar(const pcre_uchar *code, BOOL inassert)  find_firstassertedchar(const pcre_uchar *code, pcre_int32 *flags,
7638      BOOL inassert)
7639  {  {
7640  register int c = -1;  register pcre_uint32 c = 0;
7641    int cflags = REQ_NONE;
7642    
7643    *flags = REQ_NONE;
7644  do {  do {
7645     int d;     pcre_uint32 d;
7646       int dflags;
7647     int xl = (*code == OP_CBRA || *code == OP_SCBRA ||     int xl = (*code == OP_CBRA || *code == OP_SCBRA ||
7648               *code == OP_CBRAPOS || *code == OP_SCBRAPOS)? IMM2_SIZE:0;               *code == OP_CBRAPOS || *code == OP_SCBRAPOS)? IMM2_SIZE:0;
7649     const pcre_uchar *scode = first_significant_code(code + 1+LINK_SIZE + xl,     const pcre_uchar *scode = first_significant_code(code + 1+LINK_SIZE + xl,
7650       TRUE);       TRUE);
7651     register int op = *scode;     register pcre_uchar op = *scode;
7652    
7653     switch(op)     switch(op)
7654       {       {
7655       default:       default:
7656       return -1;       return 0;
7657    
7658       case OP_BRA:       case OP_BRA:
7659       case OP_BRAPOS:       case OP_BRAPOS:
# Line 7574  do { Line 7665  do {
7665       case OP_ONCE:       case OP_ONCE:
7666       case OP_ONCE_NC:       case OP_ONCE_NC:
7667       case OP_COND:       case OP_COND:
7668       if ((d = find_firstassertedchar(scode, op == OP_ASSERT)) < 0)       d = find_firstassertedchar(scode, &dflags, op == OP_ASSERT);
7669         return -1;       if (dflags < 0)
7670       if (c < 0) c = d; else if (c != d) return -1;         return 0;
7671         if (cflags < 0) { c = d; cflags = dflags; } else if (c != d || cflags != dflags) return 0;
7672       break;       break;
7673    
7674       case OP_EXACT:       case OP_EXACT:
# Line 7587  do { Line 7679  do {
7679       case OP_PLUS:       case OP_PLUS:
7680       case OP_MINPLUS:       case OP_MINPLUS:
7681       case OP_POSPLUS:       case OP_POSPLUS:
7682       if (!inassert) return -1;       if (!inassert) return 0;
7683       if (c < 0) c = scode[1];       if (cflags < 0) { c = scode[1]; cflags = 0; }
7684         else if (c != scode[1]) return -1;         else if (c != scode[1]) return 0;
7685       break;       break;
7686    
7687       case OP_EXACTI:       case OP_EXACTI:
# Line 7600  do { Line 7692  do {
7692       case OP_PLUSI:       case OP_PLUSI:
7693       case OP_MINPLUSI:       case OP_MINPLUSI:
7694       case OP_POSPLUSI:       case OP_POSPLUSI:
7695       if (!inassert) return -1;       if (!inassert) return 0;
7696       if (c < 0) c = scode[1] | REQ_CASELESS;       if (cflags < 0) { c = scode[1]; cflags = REQ_CASELESS; }
7697         else if (c != scode[1]) return -1;         else if (c != scode[1]) return 0;
7698       break;       break;
7699       }       }
7700    
7701     code += GET(code, 1);     code += GET(code, 1);
7702     }     }
7703  while (*code == OP_ALT);  while (*code == OP_ALT);
7704    
7705    *flags = cflags;
7706  return c;  return c;
7707  }  }
7708    
# Line 7676  pcre32_compile2(PCRE_SPTR32 pattern, int Line 7770  pcre32_compile2(PCRE_SPTR32 pattern, int
7770  {  {
7771  REAL_PCRE *re;  REAL_PCRE *re;
7772  int length = 1;  /* For final END opcode */  int length = 1;  /* For final END opcode */
7773  pcre_int32 firstchar, reqchar;  pcre_uint32 firstchar, reqchar;
7774    pcre_int32 firstcharflags, reqcharflags;
7775  int newline;  int newline;
7776  int errorcode = 0;  int errorcode = 0;
7777  int skipatstart = 0;  int skipatstart = 0;
# Line 7926  ptr += skipatstart; Line 8021  ptr += skipatstart;
8021  code = cworkspace;  code = cworkspace;
8022  *code = OP_BRA;  *code = OP_BRA;
8023  (void)compile_regex(cd->external_options, &code, &ptr, &errorcode, FALSE,  (void)compile_regex(cd->external_options, &code, &ptr, &errorcode, FALSE,
8024    FALSE, 0, 0, &firstchar, &reqchar, NULL, cd, &length);    FALSE, 0, 0, &firstchar, &firstcharflags, &reqchar, &reqcharflags, NULL,
8025      cd, &length);
8026  if (errorcode != 0) goto PCRE_EARLY_ERROR_RETURN;  if (errorcode != 0) goto PCRE_EARLY_ERROR_RETURN;
8027    
8028  DPRINTF(("end pre-compile: length=%d workspace=%d\n", length,  DPRINTF(("end pre-compile: length=%d workspace=%d\n", length,
# Line 8004  ptr = (const pcre_uchar *)pattern + skip Line 8100  ptr = (const pcre_uchar *)pattern + skip
8100  code = (pcre_uchar *)codestart;  code = (pcre_uchar *)codestart;
8101  *code = OP_BRA;  *code = OP_BRA;
8102  (void)compile_regex(re->options, &code, &ptr, &errorcode, FALSE, FALSE, 0, 0,  (void)compile_regex(re->options, &code, &ptr, &errorcode, FALSE, FALSE, 0, 0,
8103    &firstchar, &reqchar, NULL, cd, NULL);    &firstchar, &firstcharflags, &reqchar, &reqcharflags, NULL, cd, NULL);
8104  re->top_bracket = cd->bracount;  re->top_bracket = cd->bracount;
8105  re->top_backref = cd->top_backref;  re->top_backref = cd->top_backref;
8106  re->max_lookbehind = cd->max_lookbehind;  re->max_lookbehind = cd->max_lookbehind;
8107  re->flags = cd->external_flags | PCRE_MODE;  re->flags = cd->external_flags | PCRE_MODE;
8108    
8109  if (cd->had_accept) reqchar = REQ_NONE;   /* Must disable after (*ACCEPT) */  if (cd->had_accept)
8110      {
8111      reqchar = 0;              /* Must disable after (*ACCEPT) */
8112      reqcharflags = REQ_NONE;
8113      }
8114    
8115  /* 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. */
8116    
# Line 8060  if (errorcode == 0 && re->top_backref > Line 8160  if (errorcode == 0 && re->top_backref >
8160    
8161  /* If there were any lookbehind assertions that contained OP_RECURSE  /* If there were any lookbehind assertions that contained OP_RECURSE
8162  (recursions or subroutine calls), a flag is set for them to be checked here,  (recursions or subroutine calls), a flag is set for them to be checked here,
8163  because they may contain forward references. Actual recursions can't be fixed  because they may contain forward references. Actual recursions cannot be fixed
8164  length, but subroutine calls can. It is done like this so that those without  length, but subroutine calls can. It is done like this so that those without
8165  OP_RECURSE that are not fixed length get a diagnosic with a useful offset. The  OP_RECURSE that are not fixed length get a diagnosic with a useful offset. The
8166  exceptional ones forgo this. We scan the pattern to check that they are fixed  exceptional ones forgo this. We scan the pattern to check that they are fixed
# Line 8131  if ((re->options & PCRE_ANCHORED) == 0) Line 8231  if ((re->options & PCRE_ANCHORED) == 0)
8231    if (is_anchored(codestart, 0, cd, 0)) re->options |= PCRE_ANCHORED;    if (is_anchored(codestart, 0, cd, 0)) re->options |= PCRE_ANCHORED;
8232    else    else
8233      {      {
8234      if (firstchar < 0)      if (firstcharflags < 0)
8235        firstchar = find_firstassertedchar(codestart, FALSE);        firstchar = find_firstassertedchar(codestart, &firstcharflags, FALSE);
8236      if (firstchar >= 0)   /* Remove caseless flag for non-caseable chars */      if (firstcharflags >= 0)   /* Remove caseless flag for non-caseable chars */
8237        {        {
8238  #if defined COMPILE_PCRE8  #if defined COMPILE_PCRE8
8239        re->first_char = firstchar & 0xff;        re->first_char = firstchar & 0xff;
8240  #elif defined COMPILE_PCRE16  #elif defined COMPILE_PCRE16
8241        re->first_char = firstchar & 0xffff;        re->first_char = firstchar & 0xffff;
8242  #elif defined COMPILE_PCRE32  #elif defined COMPILE_PCRE32
8243        re->first_char = firstchar & ~REQ_MASK;        re->first_char = firstchar;
8244  #endif  #endif
8245        if ((firstchar & REQ_CASELESS) != 0)        if ((firstcharflags & REQ_CASELESS) != 0)
8246          {          {
8247  #if defined SUPPORT_UCP && !(defined COMPILE_PCRE8)  #if defined SUPPORT_UCP && !(defined COMPILE_PCRE8)
8248          /* We ignore non-ASCII first chars in 8 bit mode. */          /* We ignore non-ASCII first chars in 8 bit mode. */
# Line 8174  if ((re->options & PCRE_ANCHORED) == 0) Line 8274  if ((re->options & PCRE_ANCHORED) == 0)
8274  variable length item in the regex. Remove the caseless flag for non-caseable  variable length item in the regex. Remove the caseless flag for non-caseable
8275  bytes. */  bytes. */
8276    
8277  if (reqchar >= 0 &&  if (reqcharflags >= 0 &&
8278       ((re->options & PCRE_ANCHORED) == 0 || (reqchar & REQ_VARY) != 0))       ((re->options & PCRE_ANCHORED) == 0 || (reqcharflags & REQ_VARY) != 0))
8279    {    {
8280  #if defined COMPILE_PCRE8  #if defined COMPILE_PCRE8
8281    re->req_char = reqchar & 0xff;    re->req_char = reqchar & 0xff;
8282  #elif defined COMPILE_PCRE16  #elif defined COMPILE_PCRE16
8283    re->req_char = reqchar & 0xffff;    re->req_char = reqchar & 0xffff;
8284  #elif defined COMPILE_PCRE32  #elif defined COMPILE_PCRE32
8285    re->req_char = reqchar & ~REQ_MASK;    re->req_char = reqchar;
8286  #endif  #endif
8287    if ((reqchar & REQ_CASELESS) != 0)    if ((reqcharflags & REQ_CASELESS) != 0)
8288      {      {
8289  #if defined SUPPORT_UCP && !(defined COMPILE_PCRE8)  #if defined SUPPORT_UCP && !(defined COMPILE_PCRE8)
8290      /* We ignore non-ASCII first chars in 8 bit mode. */      /* We ignore non-ASCII first chars in 8 bit mode. */

Legend:
Removed from v.1065  
changed lines
  Added in v.1133

  ViewVC Help
Powered by ViewVC 1.1.5