pcre_compile.c

很好用的网络封装库,不熟悉网络编程的人也可以使用。使用风格良好的标准c++编写。

*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;

  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] & ctype_space) != 0;

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

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

    default:
    return FALSE;
    }

  case OP_DIGIT:
  return next == -ESC_D || next == -ESC_s || next == -ESC_W;

  case OP_NOT_DIGIT:
  return next == -ESC_d;

  case OP_WHITESPACE:
  return next == -ESC_S || next == -ESC_d || next == -ESC_w;

  case OP_NOT_WHITESPACE:
  return next == -ESC_s;

  case OP_WORDCHAR:
  return next == -ESC_W || next == -ESC_s;

  case OP_NOT_WORDCHAR:
  return next == -ESC_w || next == -ESC_d;

  default:
  return FALSE;
  }

/* Control does not reach here */
}



/*************************************************
*           Compile one branch                   *
*************************************************/

/* Scan the pattern, compiling it into the a vector. If the options are
changed during the branch, the pointer is used to change the external options
bits. This function is used during the pre-compile phase when we are trying
to find out the amount of memory needed, as well as during the real compile
phase. The value of lengthptr distinguishes the two phases.

Arguments:
  optionsptr     pointer to the option bits
  codeptr        points to the pointer to the current code point
  ptrptr         points to the current pattern pointer
  errorcodeptr   points to error code variable
  firstbyteptr   set to initial literal character, or < 0 (REQ_UNSET, REQ_NONE)
  reqbyteptr     set to the last literal character required, else < 0
  bcptr          points to current branch chain
  cd             contains pointers to tables etc.
  lengthptr      NULL during the real compile phase
                 points to length accumulator during pre-compile phase

Returns:         TRUE on success
                 FALSE, with *errorcodeptr set non-zero on error
*/

static BOOL
compile_branch(int *optionsptr, uschar **codeptr, const uschar **ptrptr,
  int *errorcodeptr, int *firstbyteptr, int *reqbyteptr, branch_chain *bcptr,
  compile_data *cd, int *lengthptr)
{
int repeat_type, op_type;
int repeat_min = 0, repeat_max = 0;      /* To please picky compilers */
int bravalue = 0;
int greedy_default, greedy_non_default;
int firstbyte, reqbyte;
int zeroreqbyte, zerofirstbyte;
int req_caseopt, reqvary, tempreqvary;
int options = *optionsptr;
int after_manual_callout = 0;
int length_prevgroup = 0;
register int c;
register uschar *code = *codeptr;
uschar *last_code = code;
uschar *orig_code = code;
uschar *tempcode;
BOOL inescq = FALSE;
BOOL groupsetfirstbyte = FALSE;
const uschar *ptr = *ptrptr;
const uschar *tempptr;
uschar *previous = NULL;
uschar *previous_callout = NULL;
uschar *save_hwm = NULL;
uschar classbits[32];

#ifdef SUPPORT_UTF8
BOOL class_utf8;
BOOL utf8 = (options & PCRE_UTF8) != 0;
uschar *class_utf8data;
uschar utf8_char[6];
#else
BOOL utf8 = FALSE;
uschar *utf8_char = NULL;
#endif

#ifdef DEBUG
if (lengthptr != NULL) DPRINTF((">> start branch\n"));
#endif

/* Set up the default and non-default settings for greediness */

greedy_default = ((options & PCRE_UNGREEDY) != 0);
greedy_non_default = greedy_default ^ 1;

/* Initialize no first byte, no required byte. REQ_UNSET means "no char
matching encountered yet". It gets changed to REQ_NONE if we hit something that
matches a non-fixed char first char; reqbyte just remains unset if we never
find one.

When we hit a repeat whose minimum is zero, we may have to adjust these values
to take the zero repeat into account. This is implemented by setting them to
zerofirstbyte and zeroreqbyte when such a repeat is encountered. The individual
item types that can be repeated set these backoff variables appropriately. */

firstbyte = reqbyte = zerofirstbyte = zeroreqbyte = REQ_UNSET;

/* The variable req_caseopt contains either the REQ_CASELESS value or zero,
according to the current setting of the caseless flag. REQ_CASELESS is a bit
value > 255. It is added into the firstbyte or reqbyte variables to record the
case status of the value. This is used only for ASCII characters. */

req_caseopt = ((options & PCRE_CASELESS) != 0)? REQ_CASELESS : 0;

/* Switch on next character until the end of the branch */

for (;; ptr++)
  {
  BOOL negate_class;
  BOOL possessive_quantifier;
  BOOL is_quantifier;
  BOOL is_recurse;
  int class_charcount;
  int class_lastchar;
  int newoptions;
  int recno;
  int skipbytes;