pcre_compile.c

SDL文件。SDL_ERROwenjian.....

/*************************************************
*          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 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 */
PUT(code, LINK_SIZE, 0);                /* Default length */
return code + 2*LINK_SIZE;
}



/*************************************************
*         Complete a callout item                *
*************************************************/

/* A callout item contains the length of the next item in the pattern, which
we can't fill in till after we have reached the relevant point. This is used
for both automatic and manual callouts.

Arguments:
  previous_callout   points to previous callout item
  ptr                current pattern pointer
  cd                 pointers to tables etc

Returns:             nothing
*/

static void
complete_callout(uschar *previous_callout, const uschar *ptr, compile_data *cd)
{
int length = ptr - cd->start_pattern - GET(previous_callout, 2);
PUT(previous_callout, 2 + LINK_SIZE, length);
}



#ifdef SUPPORT_UCP
/*************************************************
*           Get othercase range                  *
*************************************************/

/* This function is passed the start and end of a class range, in UTF-8 mode
with UCP support. It searches up the characters, looking for internal ranges of
characters in the "other" case. Each call returns the next one, updating the
start address.

Arguments:
  cptr        points to starting character value; updated
  d           end value
  ocptr       where to put start of othercase range
  odptr       where to put end of othercase range

Yield:        TRUE when range returned; FALSE when no more
*/

static BOOL
get_othercase_range(unsigned int *cptr, unsigned int d, unsigned int *ocptr,
  unsigned int *odptr)
{
unsigned int c, othercase, next;

for (c = *cptr; c <= d; c++)
  { if ((othercase = _pcre_ucp_othercase(c)) != NOTACHAR) break; }

if (c > d) return FALSE;

*ocptr = othercase;
next = othercase + 1;

for (++c; c <= d; c++)
  {
  if (_pcre_ucp_othercase(c) != next) break;
  next++;
  }

*odptr = next - 1;
*cptr = c;

return TRUE;
}
#endif  /* SUPPORT_UCP */



/*************************************************
*     Check if auto-possessifying is possible    *
*************************************************/

/* This function is called for unlimited repeats of certain items, to see
whether the next thing could possibly match the repeated item. If not, it makes
sense to automatically possessify the repeated item.

Arguments:
  op_code       the repeated op code
  this          data for this item, depends on the opcode
  utf8          TRUE in UTF-8 mode
  utf8_char     used for utf8 character bytes, NULL if not relevant
  ptr           next character in pattern
  options       options bits
  cd            contains pointers to tables etc.

Returns:        TRUE if possessifying is wanted
*/

static BOOL
check_auto_possessive(int op_code, int item, BOOL utf8, uschar *utf8_char,
  const uschar *ptr, int options, compile_data *cd)
{
int next;

/* Skip whitespace and comments in extended mode */

if ((options & PCRE_EXTENDED) != 0)
  {
  for (;;)
    {
    while ((cd->ctypes[*ptr] & ctype_space) != 0) ptr++;
    if (*ptr == '#')
      {
      while (*(++ptr) != 0)
        if (IS_NEWLINE(ptr)) { ptr += cd->nllen; break; }
      }
    else break;
    }
  }

/* If the next item is one that we can handle, get its value. A non-negative
value is a character, a negative value is an escape value. */

if (*ptr == '\\')
  {
  int temperrorcode = 0;
  next = check_escape(&ptr, &temperrorcode, cd->bracount, options, FALSE);
  if (temperrorcode != 0) return FALSE;
  ptr++;    /* Point after the escape sequence */
  }

else if ((cd->ctypes[*ptr] & ctype_meta) == 0)
  {
#ifdef SUPPORT_UTF8
  if (utf8) { GETCHARINC(next, ptr); } else
#endif
  next = *ptr++;
  }

else return FALSE;

/* Skip whitespace and comments in extended mode */

if ((options & PCRE_EXTENDED) != 0)
  {
  for (;;)
    {
    while ((cd->ctypes[*ptr] & ctype_space) != 0) ptr++;
    if (*ptr == '#')
      {
      while (*(++ptr) != 0)
        if (IS_NEWLINE(ptr)) { ptr += cd->nllen; break; }
      }
    else break;
    }
  }

/* If the next thing is itself optional, we have to give up. */

if (*ptr == '*' || *ptr == '?' || strncmp((char *)ptr, "{0,", 3) == 0)
  return FALSE;

/* Now compare the next item with the previous opcode. If the previous is a
positive single character match, "item" either contains the character or, if
"item" is greater than 127 in utf8 mode, the character's bytes are in
utf8_char. */


/* Handle cases when the next item is a character. */

if (next >= 0) switch(op_code)
  {
  case OP_CHAR:
#ifdef SUPPORT_UTF8
  if (utf8 && item > 127) { GETCHAR(item, utf8_char); }
#endif
  return item != next;

  /* For CHARNC (caseless character) we must check the other case. If we have
  Unicode property support, we can use it to test the other case of
  high-valued characters. */

  case OP_CHARNC:
#ifdef SUPPORT_UTF8
  if (utf8 && item > 127) { GETCHAR(item, utf8_char); }
#endif
  if (item == next) return FALSE;
#ifdef SUPPORT_UTF8
  if (utf8)
    {
    unsigned int othercase;
    if (next < 128) othercase = cd->fcc[next]; else
#ifdef SUPPORT_UCP
    othercase = _pcre_ucp_othercase((unsigned int)next);
#else
    othercase = NOTACHAR;
#endif
    return (unsigned int)item != othercase;
    }
  else
#endif  /* SUPPORT_UTF8 */
  return (item != cd->fcc[next]);  /* Non-UTF-8 mode */

  /* For OP_NOT, "item" must be a single-byte character. */

  case OP_NOT:
  if (next < 0) return FALSE;  /* Not a character */
  if (item == next) return TRUE;
  if ((options & PCRE_CASELESS) == 0) return FALSE;
#ifdef SUPPORT_UTF8
  if (utf8)
    {
    unsigned int othercase;
    if (next < 128) othercase = cd->fcc[next]; else
#ifdef SUPPORT_UCP
    othercase = _pcre_ucp_othercase(next);
#else
    othercase = NOTACHAR;
#endif
    return (unsigned int)item == othercase;
    }
  else
#endif  /* SUPPORT_UTF8 */
  return (item == cd->fcc[next]);  /* Non-UTF-8 mode */

  case OP_DIGIT:
  return next > 127 || (cd->ctypes[next] & ctype_digit) == 0;

  case OP_NOT_DIGIT:
  return next <= 127 && (cd->ctypes[next] & ctype_digit) != 0;

  case OP_WHITESPACE:
  return next > 127 || (cd->ctypes[next] & ctype_space) == 0;

  case OP_NOT_WHITESPACE:
  return next <= 127 && (cd->ctypes[next] & ctype_space) != 0;

  case OP_WORDCHAR:
  return next > 127 || (cd->ctypes[next] & ctype_word) == 0;

  case OP_NOT_WORDCHAR:
  return next <= 127 && (cd->ctypes[next] & ctype_word) != 0;

  case OP_HSPACE:
  case OP_NOT_HSPACE:
  switch(next)
    {
    case 0x09:
    case 0x20:
    case 0xa0:
    case 0x1680:
    case 0x180e:
    case 0x2000:
    case 0x2001:
    case 0x2002:
    case 0x2003:
    case 0x2004:
    case 0x2005:
    case 0x2006:
    case 0x2007:
    case 0x2008:
    case 0x2009:
    case 0x200A:
    case 0x202f:
    case 0x205f:
    case 0x3000:
    return op_code != OP_HSPACE;
    default:
    return op_code == OP_HSPACE;
    }

  case OP_VSPACE:
  case OP_NOT_VSPACE:
  switch(next)
    {
    case 0x0a:
    case 0x0b:
    case 0x0c:
    case 0x0d:
    case 0x85:
    case 0x2028:
    case 0x2029:
    return op_code != OP_VSPACE;
    default:
    return op_code == OP_VSPACE;
    }

  default:
  return FALSE;
  }


/* Handle the case when the next item is \d, \s, etc. */

switch(op_code)
  {
  case OP_CHAR:
  case OP_CHARNC:
#ifdef SUPPORT_UTF8
  if (utf8 && item > 127) { GETCHAR(item, utf8_char); }
#endif
  switch(-next)
    {
    case ESC_d:
    return item > 127 || (cd->ctypes[item] & ctype_digit) == 0;

    case ESC_D:
    return item <= 127 && (cd->ctypes[item] & ctype_digit) != 0;

    case ESC_s:
    return item > 127 || (cd->ctypes[item] & ctype_space) == 0;

    case ESC_S:
    return item <= 127 && (cd->ctypes[item] & cty