fnmatch_loop.c

Intercom 是一个 Unix系统上灵活的语音传输软件。支持标准音频压缩比如GSM, G.711, and G.72x和其他音频编码。Intercom专为高速网络设计来传输高品质的语音

			if (!is_range && c == fn)
			  goto matched;

			cold = c;
			c = *p++;
		      }

		    if (c == L('-') && *p != L(']'))
		      {
#if _LIBC
			/* We have to find the collation sequence
			   value for C.  Collation sequence is nothing
			   we can regularly access.  The sequence
			   value is defined by the order in which the
			   definitions of the collation values for the
			   various characters appear in the source
			   file.  A strange concept, nowhere
			   documented.  */
			uint32_t fcollseq;
			uint32_t lcollseq;
			UCHAR cend = *p++;

# ifdef WIDE_CHAR_VERSION
			/* Search in the `names' array for the characters.  */
			fcollseq = collseq_table_lookup (collseq, fn);
			if (fcollseq == ~((uint32_t) 0))
			  /* XXX We don't know anything about the character
			     we are supposed to match.  This means we are
			     failing.  */
			  goto range_not_matched;

			if (is_seqval)
			  lcollseq = cold;
			else
			  lcollseq = collseq_table_lookup (collseq, cold);
# else
			fcollseq = collseq[fn];
			lcollseq = is_seqval ? cold : collseq[(UCHAR) cold];
# endif

			is_seqval = 0;
			if (cend == L('[') && *p == L('.'))
			  {
			    uint32_t nrules =
			      _NL_CURRENT_WORD (LC_COLLATE,
						_NL_COLLATE_NRULES);
			    const CHAR *startp = p;
			    size_t c1 = 0;

			    while (1)
			      {
				c = *++p;
				if (c == L('.') && p[1] == L(']'))
				  {
				    p += 2;
				    break;
				  }
				if (c == '\0')
				  return FNM_NOMATCH;
				++c1;
			      }

			    if (nrules == 0)
			      {
				/* There are no names defined in the
				   collation data.  Therefore we only
				   accept the trivial names consisting
				   of the character itself.  */
				if (c1 != 1)
				  return FNM_NOMATCH;

				cend = startp[1];
			      }
			    else
			      {
				int32_t table_size;
				const int32_t *symb_table;
# ifdef WIDE_CHAR_VERSION
				char str[c1];
				unsigned int strcnt;
# else
#  define str (startp + 1)
# endif
				const unsigned char *extra;
				int32_t idx;
				int32_t elem;
				int32_t second;
				int32_t hash;

# ifdef WIDE_CHAR_VERSION
				/* We have to convert the name to a single-byte
				   string.  This is possible since the names
				   consist of ASCII characters and the internal
				   representation is UCS4.  */
				for (strcnt = 0; strcnt < c1; ++strcnt)
				  str[strcnt] = startp[1 + strcnt];
# endif

				table_size =
				  _NL_CURRENT_WORD (LC_COLLATE,
						    _NL_COLLATE_SYMB_HASH_SIZEMB);
				symb_table = (const int32_t *)
				  _NL_CURRENT (LC_COLLATE,
					       _NL_COLLATE_SYMB_TABLEMB);
				extra = (const unsigned char *)
				  _NL_CURRENT (LC_COLLATE,
					       _NL_COLLATE_SYMB_EXTRAMB);

				/* Locate the character in the hashing
                                   table.  */
				hash = elem_hash (str, c1);

				idx = 0;
				elem = hash % table_size;
				second = hash % (table_size - 2);
				while (symb_table[2 * elem] != 0)
				  {
				/* First compare the hashing value.  */
				    if (symb_table[2 * elem] == hash
					&& (c1
					    == extra[symb_table[2 * elem + 1]])
					&& memcmp (str,
						   &extra[symb_table[2 * elem + 1]
							 + 1], c1) == 0)
				      {
					/* Yep, this is the entry.  */
					idx = symb_table[2 * elem + 1];
					idx += 1 + extra[idx];
					break;
				      }

				    /* Next entry.  */
				    elem += second;
				  }

				if (symb_table[2 * elem] != 0)
				  {
				    /* Compare the byte sequence but only if
				       this is not part of a range.  */
# ifdef WIDE_CHAR_VERSION
				    int32_t *wextra;

				    idx += 1 + extra[idx];
				    /* Adjust for the alignment.  */
				    idx = (idx + 3) & ~4;

				    wextra = (int32_t *) &extra[idx + 4];
# endif
				    /* Get the collation sequence value.  */
				    is_seqval = 1;
# ifdef WIDE_CHAR_VERSION
				    cend = wextra[1 + wextra[idx]];
# else
				    /* Adjust for the alignment.  */
				    idx += 1 + extra[idx];
				    idx = (idx + 3) & ~4;
				    cend = *((int32_t *) &extra[idx]);
# endif
				  }
				else if (symb_table[2 * elem] != 0 && c1 == 1)
				  {
				    cend = str[0];
				    c = *p++;
				  }
				else
				  return FNM_NOMATCH;
			      }
# undef str
			  }
			else
			  {
			    if (!(flags & FNM_NOESCAPE) && cend == L('\\'))
			      cend = *p++;
			    if (cend == L('\0'))
			      return FNM_NOMATCH;
			    cend = FOLD (cend);
			  }

			/* XXX It is not entirely clear to me how to handle
			   characters which are not mentioned in the
			   collation specification.  */
			if (
# ifdef WIDE_CHAR_VERSION
			    lcollseq == 0xffffffff ||
# endif
			    lcollseq <= fcollseq)
			  {
			    /* We have to look at the upper bound.  */
			    uint32_t hcollseq;

			    if (is_seqval)
			      hcollseq = cend;
			    else
			      {
# ifdef WIDE_CHAR_VERSION
				hcollseq =
				  collseq_table_lookup (collseq, cend);
				if (hcollseq == ~((uint32_t) 0))
				  {
				    /* Hum, no information about the upper
				       bound.  The matching succeeds if the
				       lower bound is matched exactly.  */
				    if (lcollseq != fcollseq)
				      goto range_not_matched;

				    goto matched;
				  }
# else
				hcollseq = collseq[cend];
# endif
			      }

			    if (lcollseq <= hcollseq && fcollseq <= hcollseq)
			      goto matched;
			  }
# ifdef WIDE_CHAR_VERSION
		      range_not_matched:
# endif
#else
			/* We use a boring value comparison of the character
			   values.  This is better than comparing using
			   `strcoll' since the latter would have surprising
			   and sometimes fatal consequences.  */
			UCHAR cend = *p++;

			if (!(flags & FNM_NOESCAPE) && cend == L('\\'))
			  cend = *p++;
			if (cend == L('\0'))
			  return FNM_NOMATCH;

			/* It is a range.  */
			if (cold <= fn && fn <= cend)
			  goto matched;
#endif

			c = *p++;
		      }
		  }

		if (c == L(']'))
		  break;
	      }

	    if (!not)
	      return FNM_NOMATCH;
	    break;

	  matched:
	    /* Skip the rest of the [...] that already matched.  */
	    do
	      {
	      ignore_next:
		c = *p++;

		if (c == L('\0'))
		  /* [... (unterminated) loses.  */
		  return FNM_NOMATCH;

		if (!(flags & FNM_NOESCAPE) && c == L('\\'))
		  {
		    if (*p == L('\0'))
		      return FNM_NOMATCH;
		    /* XXX 1003.2d11 is unclear if this is right.  */
		    ++p;
		  }
		else if (c == L('[') && *p == L(':'))
		  {
		    int c1 = 0;
		    const CHAR *startp = p;

		    while (1)
		      {
			c = *++p;
			if (++c1 == CHAR_CLASS_MAX_LENGTH)
			  return FNM_NOMATCH;

			if (*p == L(':') && p[1] == L(']'))
			  break;

			if (c < L('a') || c >= L('z'))
			  {
			    p = startp;
			    goto ignore_next;
			  }
		      }
		    p += 2;
		    c = *p++;
		  }
		else if (c == L('[') && *p == L('='))
		  {
		    c = *++p;
		    if (c == L('\0'))
		      return FNM_NOMATCH;
		    c = *++p;
		    if (c != L('=') || p[1] != L(']'))
		      return FNM_NOMATCH;
		    p += 2;
		    c = *p++;
		  }
		else if (c == L('[') && *p == L('.'))
		  {
		    ++p;
		    while (1)
		      {
			c = *++p;
			if (c == '\0')
			  return FNM_NOMATCH;

			if (*p == L('.') && p[1] == L(']'))
			  break;
		      }
		    p += 2;
		    c = *p++;
		  }
	      }
	    while (c != L(']'));
	    if (not)
	      return FNM_NOMATCH;
	  }
	  break;

	case L('+'):
	case L('@'):
	case L('!'):
	  if (__builtin_expect (flags & FNM_EXTMATCH, 0) && *p == '(')
	    {
	      int res;

	      res = EXT (c, p, n, string_end, no_leading_period, flags);
	      if (res != -1)
		return res;
	    }
	  goto normal_match;

	case L('/'):
	  if (NO_LEADING_PERIOD (flags))
	    {
	      if (n == string_end || c != *n)
		return FNM_NOMATCH;

	      new_no_leading_period = 1;
	      break;
	    }
	  /* FALLTHROUGH */
	default:
	normal_match:
	  if (n == string_end || c != FOLD ((UCHAR) *n))
	    return FNM_NOMATCH;
	}

      no_leading_period = new_no_leading_period;
      ++n;
    }

  if (n == string_end)
    return 0;

  if ((flags & FNM_LEADING_DIR) && n != string_end && *n == L('/'))
    /* The FNM_LEADING_DIR flag says that "foo*" matches "foobar/frobozz".  */
    return 0;

  return FNM_NOMATCH;
}


static const CHAR *
internal_function
END (const CHAR *pattern)
{
  const CHAR *p = pattern;

  while (1)
    if (*++p == L('\0'))
      /* This is an invalid pattern.  */
      return pattern;
    else if (*p == L('['))
      {
	/* Handle brackets special.  */
	if (posixly_correct == 0)
	  posixly_correct = getenv ("POSIXLY_CORRECT") != NULL ? 1 : -1;

	/* Skip the not sign.  We have to recognize it because of a possibly
	   following ']'.  */
	if (*++p == L('!') || (posixly_correct < 0 && *p == L('^')))
	  ++p;
	/* A leading ']' is recognized as such.  */
	if (*p == L(']'))
	  ++p;
	/* Skip over all characters of the list.  */
	while (*p != L(']'))
	  if (*p++ == L('\0'))
	    /* This is no valid pattern.  */
	    return pattern;
      }
    else if ((*p == L('?') || *p == L('*') || *p == L('+') || *p == L('@')
	      || *p == L('!')) && p[1] == L('('))
      p = END (p + 1);
    else if (*p == L(')'))
      break;

  return p + 1;
}


static int
internal_function
EXT (INT opt, const CHAR *pattern, const CHAR *string, const CHAR *string_end,
     int no_leading_period, int flags)
{
  const CHAR *startp;
  int level;
  struct patternlist
  {
    struct patternlist *next;
    CHAR str[0];
  } *list = NULL;
  struct patternlist **lastp = &list;
  size_t pattern_len = STRLEN (pattern);
  const CHAR *p;
  const CHAR *rs;

  /* Parse the pattern.  Store the individual parts in the list.  */
  level = 0;
  for (startp = p = pattern + 1; level >= 0; ++p)
    if (*p == L('\0'))
      /* This is an invalid pattern.  */
      return -1;
    else if (*p == L('['))
      {
	/* Handle brackets special.  */
	if (posixly_correct == 0)
	  posixly_correct = getenv ("POSIXLY_CORRECT") != NULL ? 1 : -1;

	/* Skip the not sign.  We have to recognize it because of a possibly
	   following ']'.  */
	if (*++p == L('!') || (posixly_correct < 0 && *p == L('^')))
	  ++p;
	/* A leading ']' is recognized as such.  */
	if (*p == L(']'))
	  ++p;
	/* Skip over all characters of the list.  */
	while (*p != L(']'))
	  if (*p++ == L('\0'))
	    /* This is no valid pattern.  */
	    return -1;
      }
    else if ((*p == L('?') || *p == L('*') || *p == L('+') || *p == L('@')
	      || *p == L('!')) && p[1] == L('('))
      /* Remember the nesting level.  */
      ++level;
    else if (*p == L(')'))
      {
	if (level-- == 0)
	  {
	    /* This means we found the end of the pattern.  */
#define NEW_PATTERN \
	    struct patternlist *newp;					      \
									      \
	    if (opt == L('?') || opt == L('@'))				      \
	      newp = alloca (sizeof (struct patternlist)		      \
			     + (pattern_len * sizeof (CHAR)));		      \
	    else							      \
	      newp = alloca (sizeof (struct patternlist)		      \
			     + ((p - startp + 1) * sizeof (CHAR)));	      \
	    *((CHAR *) MEMPCPY (newp->str, startp, p - startp)) = L('\0');    \
	    newp->next = NULL;						      \
	    *lastp = newp;						      \
	    lastp = &newp->next
	    NEW_PATTERN;
	  }
      }
    else if (*p == L('|'))
      {
	if (level == 0)
	  {
	    NEW_PATTERN;
	    startp = p + 1;
	  }
      }
  assert (list != NULL);
  assert (p[-1] == L(')'));
#undef NEW_PATTERN

  switch (opt)
    {
    case L('*'):
      if (FCT (p, string, string_end, no_leading_period, flags) == 0)
	return 0;
      /* FALLTHROUGH */

    case L('+'):
      do
	{
	  for (rs = string; rs <= string_end; ++rs)
	    /* First match the prefix with the current pattern with the
	       current pattern.  */
	    if (FCT (list->str, string, rs, no_leading_period,
		     flags & FNM_FILE_NAME ? flags : flags & ~FNM_PERIOD) == 0
		/* This was successful.  Now match the rest with the rest
		   of the pattern.  */
		&& (FCT (p, rs, string_end,
			 rs == string
			 ? no_leading_period
			 : rs[-1] == '/' && NO_LEADING_PERIOD (flags) ? 1 : 0,
			 flags & FNM_FILE_NAME
			 ? flags : flags & ~FNM_PERIOD) == 0
		    /* This didn't work.  Try the whole pattern.  */
		    || (rs != string
			&& FCT (pattern - 1, rs, string_end,
				rs == string
				? no_leading_period
				: (rs[-1] == '/' && NO_LEADING_PERIOD (flags)
				   ? 1 : 0),
				flags & FNM_FILE_NAME
				? flags : flags & ~FNM_PERIOD) == 0)))
	      /* It worked.  Signal success.  */
	      return 0;
	}
      while ((list = list->next) != NULL);

      /* None of the patterns lead to a match.  */
      return FNM_NOMATCH;

    case L('?'):
      if (FCT (p, string, string_end, no_leading_period, flags) == 0)
	return 0;
      /* FALLTHROUGH */

    case L('@'):
      do
	/* I cannot believe it but `strcat' is actually acceptable
	   here.  Match the entire string with the prefix from the
	   pattern list and the rest of the pattern following the
	   pattern list.  */
	if (FCT (STRCAT (list->str, p), string, string_end,
		 no_leading_period,
		 flags & FNM_FILE_NAME ? flags : flags & ~FNM_PERIOD) == 0)
	  /* It worked.  Signal success.  */
	  return 0;
      while ((list = list->next) != NULL);

      /* None of the patterns lead to a match.  */
      return FNM_NOMATCH;

    case L('!'):
      for (rs = string; rs <= string_end; ++rs)
	{
	  struct patternlist *runp;

	  for (runp = list; runp != NULL; runp = runp->next)
	    if (FCT (runp->str, string, rs,  no_leading_period,
		     flags & FNM_FILE_NAME ? flags : flags & ~FNM_PERIOD) == 0)
	      break;

	  /* If none of the patterns matched see whether the rest does.  */
	  if (runp == NULL
	      && (FCT (p, rs, string_end,
		       rs == string
		       ? no_leading_period
		       : rs[-1] == '/' && NO_LEADING_PERIOD (flags) ? 1 : 0,
		       flags & FNM_FILE_NAME ? flags : flags & ~FNM_PERIOD)
		  == 0))
	    /* This is successful.  */
	    return 0;
	}

      /* None of the patterns together with the rest of the pattern
	 lead to a match.  */
      return FNM_NOMATCH;

    default:
      assert (! "Invalid extended matching operator");
      break;
    }

  return -1;
}


#undef FOLD
#undef CHAR
#undef UCHAR
#undef INT
#undef FCT
#undef EXT
#undef END
#undef MEMPCPY
#undef MEMCHR
#undef STRCOLL
#undef STRLEN
#undef STRCAT
#undef L
#undef BTOWC