pcre_compile.c

this is a glib for c language

      case OP_TYPEMINQUERY:
      case OP_TYPEPOSSTAR:
      case OP_TYPEPOSPLUS:
      case OP_TYPEPOSQUERY:
      if (code[1] == OP_PROP || code[1] == OP_NOTPROP) code += 2;
      break;

      case OP_TYPEPOSUPTO:
      case OP_TYPEUPTO:
      case OP_TYPEMINUPTO:
      case OP_TYPEEXACT:
      if (code[3] == OP_PROP || code[3] == OP_NOTPROP) code += 2;
      break;
      }

    /* Add in the fixed length from the table */

    code += _pcre_OP_lengths[c];

    /* In UTF-8 mode, opcodes that are followed by a character may be followed
    by a multi-byte character. The length in the table is a minimum, so we have
    to arrange to skip the extra bytes. */

#ifdef SUPPORT_UTF8
    if (utf8) switch(c)
      {
      case OP_CHAR:
      case OP_CHARNC:
      case OP_EXACT:
      case OP_UPTO:
      case OP_MINUPTO:
      case OP_POSUPTO:
      case OP_STAR:
      case OP_MINSTAR:
      case OP_POSSTAR:
      case OP_PLUS:
      case OP_MINPLUS:
      case OP_POSPLUS:
      case OP_QUERY:
      case OP_MINQUERY:
      case OP_POSQUERY:
      if (code[-1] >= 0xc0) code += _pcre_utf8_table4[code[-1] & 0x3f];
      break;
      }
#endif
    }
  }
}



/*************************************************
*    Scan compiled branch for non-emptiness      *
*************************************************/

/* This function scans through a branch of a compiled pattern to see whether it
can match the empty string or not. It is called from could_be_empty()
below and from compile_branch() when checking for an unlimited repeat of a
group that can match nothing. Note that first_significant_code() skips over
backward and negative forward assertions when its final argument is TRUE. If we
hit an unclosed bracket, we return "empty" - this means we've struck an inner
bracket whose current branch will already have been scanned.

Arguments:
  code        points to start of search
  endcode     points to where to stop
  utf8        TRUE if in UTF8 mode

Returns:      TRUE if what is matched could be empty
*/

static BOOL
could_be_empty_branch(const uschar *code, const uschar *endcode, BOOL utf8)
{
register int c;
for (code = first_significant_code(code + _pcre_OP_lengths[*code], NULL, 0, TRUE);
     code < endcode;
     code = first_significant_code(code + _pcre_OP_lengths[c], NULL, 0, TRUE))
  {
  const uschar *ccode;

  c = *code;

  /* Skip over forward assertions; the other assertions are skipped by
  first_significant_code() with a TRUE final argument. */

  if (c == OP_ASSERT)
    {
    do code += GET(code, 1); while (*code == OP_ALT);
    c = *code;
    continue;
    }

  /* Groups with zero repeats can of course be empty; skip them. */

  if (c == OP_BRAZERO || c == OP_BRAMINZERO || c == OP_SKIPZERO)
    {
    code += _pcre_OP_lengths[c];
    do code += GET(code, 1); while (*code == OP_ALT);
    c = *code;
    continue;
    }

  /* For other groups, scan the branches. */

  if (c == OP_BRA || c == OP_CBRA || c == OP_ONCE || c == OP_COND)
    {
    BOOL empty_branch;
    if (GET(code, 1) == 0) return TRUE;    /* Hit unclosed bracket */

    /* Scan a closed bracket */

    empty_branch = FALSE;
    do
      {
      if (!empty_branch && could_be_empty_branch(code, endcode, utf8))
        empty_branch = TRUE;
      code += GET(code, 1);
      }
    while (*code == OP_ALT);
    if (!empty_branch) return FALSE;   /* All branches are non-empty */
    c = *code;
    continue;
    }

  /* Handle the other opcodes */

  switch (c)
    {
    /* Check for quantifiers after a class. XCLASS is used for classes that
    cannot be represented just by a bit map. This includes negated single
    high-valued characters. The length in _pcre_OP_lengths[] is zero; the
    actual length is stored in the compiled code, so we must update "code"
    here. */

#ifdef SUPPORT_UTF8
    case OP_XCLASS:
    ccode = code += GET(code, 1);
    goto CHECK_CLASS_REPEAT;
#endif

    case OP_CLASS:
    case OP_NCLASS:
    ccode = code + 33;

#ifdef SUPPORT_UTF8
    CHECK_CLASS_REPEAT:
#endif

    switch (*ccode)
      {
      case OP_CRSTAR:            /* These could be empty; continue */
      case OP_CRMINSTAR:
      case OP_CRQUERY:
      case OP_CRMINQUERY:
      break;

      default:                   /* Non-repeat => class must match */
      case OP_CRPLUS:            /* These repeats aren't empty */
      case OP_CRMINPLUS:
      return FALSE;

      case OP_CRRANGE:
      case OP_CRMINRANGE:
      if (GET2(ccode, 1) > 0) return FALSE;  /* Minimum > 0 */
      break;
      }
    break;

    /* Opcodes that must match a character */

    case OP_PROP:
    case OP_NOTPROP:
    case OP_EXTUNI:
    case OP_NOT_DIGIT:
    case OP_DIGIT:
    case OP_NOT_WHITESPACE:
    case OP_WHITESPACE:
    case OP_NOT_WORDCHAR:
    case OP_WORDCHAR:
    case OP_ANY:
    case OP_ALLANY:
    case OP_ANYBYTE:
    case OP_CHAR:
    case OP_CHARNC:
    case OP_NOT:
    case OP_PLUS:
    case OP_MINPLUS:
    case OP_POSPLUS:
    case OP_EXACT:
    case OP_NOTPLUS:
    case OP_NOTMINPLUS:
    case OP_NOTPOSPLUS:
    case OP_NOTEXACT:
    case OP_TYPEPLUS:
    case OP_TYPEMINPLUS:
    case OP_TYPEPOSPLUS:
    case OP_TYPEEXACT:
    return FALSE;

    /* These are going to continue, as they may be empty, but we have to
    fudge the length for the \p and \P cases. */

    case OP_TYPESTAR:
    case OP_TYPEMINSTAR:
    case OP_TYPEPOSSTAR:
    case OP_TYPEQUERY:
    case OP_TYPEMINQUERY:
    case OP_TYPEPOSQUERY:
    if (code[1] == OP_PROP || code[1] == OP_NOTPROP) code += 2;
    break;

    /* Same for these */

    case OP_TYPEUPTO:
    case OP_TYPEMINUPTO:
    case OP_TYPEPOSUPTO:
    if (code[3] == OP_PROP || code[3] == OP_NOTPROP) code += 2;
    break;

    /* End of branch */

    case OP_KET:
    case OP_KETRMAX:
    case OP_KETRMIN:
    case OP_ALT:
    return TRUE;

    /* In UTF-8 mode, STAR, MINSTAR, POSSTAR, QUERY, MINQUERY, POSQUERY, UPTO,
    MINUPTO, and POSUPTO may be followed by a multibyte character */

#ifdef SUPPORT_UTF8
    case OP_STAR:
    case OP_MINSTAR:
    case OP_POSSTAR:
    case OP_QUERY:
    case OP_MINQUERY:
    case OP_POSQUERY:
    case OP_UPTO:
    case OP_MINUPTO:
    case OP_POSUPTO:
    if (utf8) while ((code[2] & 0xc0) == 0x80) code++;
    break;
#endif
    }
  }

return TRUE;
}



/*************************************************
*    Scan compiled regex for non-emptiness       *
*************************************************/

/* This function is called to check for left recursive calls. We want to check
the current branch of the current pattern to see if it could match the empty
string. If it could, we must look outwards for branches at other levels,
stopping when we pass beyond the bracket which is the subject of the recursion.

Arguments:
  code        points to start of the recursion
  endcode     points to where to stop (current RECURSE item)
  bcptr       points to the chain of current (unclosed) branch starts
  utf8        TRUE if in UTF-8 mode

Returns:      TRUE if what is matched could be empty
*/

static BOOL
could_be_empty(const uschar *code, const uschar *endcode, branch_chain *bcptr,
  BOOL utf8)
{
while (bcptr != NULL && bcptr->current >= code)
  {
  if (!could_be_empty_branch(bcptr->current, endcode, utf8)) return FALSE;
  bcptr = bcptr->outer;
  }
return TRUE;
}



/*************************************************
*           Check for POSIX class syntax         *
*************************************************/

/* This function is called when the sequence "[:" or "[." or "[=" is
encountered in a character class. It checks whether this is followed by a
sequence of characters terminated by a matching ":]" or ".]" or "=]". If we
reach an unescaped ']' without the special preceding character, return FALSE.

Originally, this function only recognized a sequence of letters between the
terminators, but it seems that Perl recognizes any sequence of characters,
though of course unknown POSIX names are subsequently rejected. Perl gives an
"Unknown POSIX class" error for [:f\oo:] for example, where previously PCRE
didn't consider this to be a POSIX class. Likewise for [:1234:].

The problem in trying to be exactly like Perl is in the handling of escapes. We
have to be sure that [abc[:x\]pqr] is *not* treated as containing a POSIX
class, but [abc[:x\]pqr:]] is (so that an error can be generated). The code
below handles the special case of \], but does not try to do any other escape
processing. This makes it different from Perl for cases such as [:l\ower:]
where Perl recognizes it as the POSIX class "lower" but PCRE does not recognize
"l\ower". This is a lesser evil that not diagnosing bad classes when Perl does,
I think.

Arguments:
  ptr      pointer to the initial [
  endptr   where to return the end pointer

Returns:   TRUE or FALSE
*/

static BOOL
check_posix_syntax(const uschar *ptr, const uschar **endptr)
{
int terminator;          /* Don't combine these lines; the Solaris cc */
terminator = *(++ptr);   /* compiler warns about "non-constant" initializer. */
for (++ptr; *ptr != 0; ptr++)
  {
  if (*ptr == '\\' && ptr[1] == ']') ptr++; else
    {
    if (*ptr == ']') return FALSE;
    if (*ptr == terminator && ptr[1] == ']')
      {
      *endptr = ptr;
      return TRUE;
      }
    }
  }
return FALSE;
}




/*************************************************
*          Check POSIX class name                *
*************************************************/

/* This function is called to check the name given in a POSIX-style class entry
such as [:alnum:].

Arguments:
  ptr        points to the first letter
  len        the length of the name

Returns:     a value representing the name, or -1 if unknown
*/

static int
check_posix_name(const uschar *ptr, int len)
{
const char *pn = posix_names;
register int yield = 0;
while (posix_name_lengths[yield] != 0)
  {
  if (len == posix_name_lengths[yield] &&
    strncmp((const char *)ptr, pn, len) == 0) return yield;
  pn += posix_name_lengths[yield] + 1;
  yield++;
  }
return -1;
}


/*************************************************
*    Adjust OP_RECURSE items in repeated group   *
*************************************************/

/* OP_RECURSE items contain an offset from the start of the regex to the group
that is referenced. This means that groups can be replicated for fixed
repetition simply by copying (because the recursion is allowed to refer to
earlier groups that are outside the current group). However, when a group is
optional (i.e. the minimum quantifier is zero), OP_BRAZERO or OP_SKIPZERO is
inserted before it, after it has been compiled. This means that any OP_RECURSE
items within it that refer to the group itself or any contained groups have to
have their offsets adjusted. That one of the jobs of this function. Before it
is called, the partially compiled regex must be temporarily terminated with
OP_END.

This function has been extended with the possibility of forward references for
recursions and subroutine calls. It must also check the list of such references
for the group we are dealing with. If it finds that one of the recursions in
the current group is on this list, it adjusts the offset in the list, not the
value in the reference (which is a group number).

Arguments:
  group      points to the start of the group
  adjust     the amount by which the group is to be moved
  utf8       TRUE in UTF-8 mode
  cd         contains pointers to tables etc.
  save_hwm   the hwm forward reference pointer at the start of the group

Returns:     nothing
*/

static void
adjust_recurse(uschar *group, int adjust, BOOL utf8, compile_data *cd,
  uschar *save_hwm)
{
uschar *ptr = group;

while ((ptr = (uschar *)find_recurse(ptr, utf8)) != NULL)
  {
  int offset;
  uschar *hc;

  /* See if this recursion is on the forward reference list. If so, adjust the
  reference. */

  for (hc = save_hwm; hc < cd->hwm; hc += LINK_SIZE)
    {
    offset = GET(hc, 0);
    if (cd->start_code + offset == ptr + 1)
      {
      PUT(hc, 0, offset + adjust);
      break;
      }
    }

  /* Otherwise, adjust the recursion offset if it's after the start of this
  group. */

  if (hc >= cd->hwm)
    {
    offset = GET(ptr, 1);
    if (cd->start_code + offset >= group) PUT(ptr, 1, offset + adjust);
    }

  ptr += 1 + LINK_SIZE;
  }
}



/*************************************************
*        Insert an automatic callout point       *
*************************************************/

/* This function is called when the PCRE_AUTO_CALLOUT option is set, to insert
callout points before each pattern item.

Arguments:
  code           current code pointer
  ptr            current pattern pointer
  cd             pointers to tables etc

Returns:         new code pointer
*/

static uschar *
auto_callout(uschar *code, const uschar *ptr, compile_data *cd)
{
*code++ = OP_CALLOUT;
*code++ = 255;
PUT(code, 0, ptr - cd->start_pattern);  /* Pattern offset */