Beispiel #1
0
static int
quoted_file_exists (const char* pz_src_path,
                    const char* pz_file_path, 
                    const char* pz_file)
{
  char z[ MAXPATHLEN ];
  char* pz;
  sprintf (z, "%s/%s/", pz_src_path, pz_file_path);
  pz = z + strlen ( z );

  for (;;) {
    char ch = *pz_file++;
    if (! ISGRAPH( ch ))
      return 0;
    if (ch == '"')
      break;
    *pz++ = ch;
  }
  *pz = '\0';
  {
    struct stat s;
    if (stat (z, &s) != 0)
      return 0;
    return S_ISREG( s.st_mode );
  }
}
Beispiel #2
0
static int get_last_char_offset(char *fat_string)
{
    int i = 0;

    for (i = strlen(fat_string) - 1; i >= 0; i--) {
        if (ISGRAPH((int) *(fat_string + i)))
            return (i + 1);
    }

    return (strlen(fat_string));
}
Beispiel #3
0
static void lograw(unsigned char *buffer, ssize_t len)
{
  char data[120];
  ssize_t i;
  unsigned char *ptr = buffer;
  char *optr = data;
  ssize_t width = 0;
  int left = sizeof(data);

  for(i = 0; i<len; i++) {
    switch(ptr[i]) {
    case '\n':
      msnprintf(optr, left, "\\n");
      width += 2;
      optr += 2;
      left -= 2;
      break;
    case '\r':
      msnprintf(optr, left, "\\r");
      width += 2;
      optr += 2;
      left -= 2;
      break;
    default:
      msnprintf(optr, left, "%c", (ISGRAPH(ptr[i]) ||
                                   ptr[i] == 0x20) ?ptr[i]:'.');
      width++;
      optr++;
      left--;
      break;
    }

    if(width>60) {
      logmsg("'%s'", data);
      width = 0;
      optr = data;
      left = sizeof(data);
    }
  }
  if(width)
    logmsg("'%s'", data);
}
Beispiel #4
0
static int loop(const unsigned char *pattern, const unsigned char *string)
{
  loop_state state = CURLFNM_LOOP_DEFAULT;
  unsigned char *p = (unsigned char *)pattern;
  unsigned char *s = (unsigned char *)string;
  unsigned char charset[CURLFNM_CHSET_SIZE] = { 0 };
  int rc = 0;

  for(;;) {
    switch(state) {
    case CURLFNM_LOOP_DEFAULT:
      if(*p == '*') {
        while(*(p+1) == '*') /* eliminate multiple stars */
          p++;
        if(*s == '\0' && *(p+1) == '\0')
          return CURL_FNMATCH_MATCH;
        rc = loop(p + 1, s); /* *.txt matches .txt <=> .txt matches .txt */
        if(rc == CURL_FNMATCH_MATCH)
          return CURL_FNMATCH_MATCH;
        if(*s) /* let the star eat up one character */
          s++;
        else
          return CURL_FNMATCH_NOMATCH;
      }
      else if(*p == '?') {
        if(ISPRINT(*s)) {
          s++;
          p++;
        }
        else if(*s == '\0')
          return CURL_FNMATCH_NOMATCH;
        else
          return CURL_FNMATCH_FAIL; /* cannot deal with other character */
      }
      else if(*p == '\0') {
        if(*s == '\0')
          return CURL_FNMATCH_MATCH;
        else
          return CURL_FNMATCH_NOMATCH;
      }
      else if(*p == '\\') {
        state = CURLFNM_LOOP_BACKSLASH;
        p++;
      }
      else if(*p == '[') {
        unsigned char *pp = p+1; /* cannot handle with pointer to register */
        if(setcharset(&pp, charset)) {
          int found = FALSE;
          if(charset[(unsigned int)*s])
            found = TRUE;
          else if(charset[CURLFNM_ALNUM])
            found = ISALNUM(*s);
          else if(charset[CURLFNM_ALPHA])
            found = ISALPHA(*s);
          else if(charset[CURLFNM_DIGIT])
            found = ISDIGIT(*s);
          else if(charset[CURLFNM_XDIGIT])
            found = ISXDIGIT(*s);
          else if(charset[CURLFNM_PRINT])
            found = ISPRINT(*s);
          else if(charset[CURLFNM_SPACE])
            found = ISSPACE(*s);
          else if(charset[CURLFNM_UPPER])
            found = ISUPPER(*s);
          else if(charset[CURLFNM_LOWER])
            found = ISLOWER(*s);
          else if(charset[CURLFNM_BLANK])
            found = ISBLANK(*s);
          else if(charset[CURLFNM_GRAPH])
            found = ISGRAPH(*s);

          if(charset[CURLFNM_NEGATE])
            found = !found;

          if(found) {
            p = pp+1;
            s++;
            memset(charset, 0, CURLFNM_CHSET_SIZE);
          }
          else
            return CURL_FNMATCH_NOMATCH;
        }
        else
          return CURL_FNMATCH_FAIL;
      }
      else {
        if(*p++ != *s++)
          return CURL_FNMATCH_NOMATCH;
      }
      break;
    case CURLFNM_LOOP_BACKSLASH:
      if(ISPRINT(*p)) {
        if(*p++ == *s++)
          state = CURLFNM_LOOP_DEFAULT;
        else
          return CURL_FNMATCH_NOMATCH;
      }
      else
        return CURL_FNMATCH_FAIL;
      break;
    }
  }
}
Beispiel #5
0
enum cpp_ttype
c_lex_with_flags (tree *value, location_t *loc, unsigned char *cpp_flags)
{
  static bool no_more_pch;
  const cpp_token *tok;
  enum cpp_ttype type;

  timevar_push (TV_CPP);
 retry:
  tok = cpp_get_token (parse_in);
  type = tok->type;
  
 retry_after_at:
#ifdef USE_MAPPED_LOCATION
  *loc = tok->src_loc;
#else
  *loc = input_location;
#endif
  switch (type)
    {
    case CPP_PADDING:
      goto retry;
      
    case CPP_NAME:
      *value = HT_IDENT_TO_GCC_IDENT (HT_NODE (tok->val.node));
      break;

    case CPP_NUMBER:
      {
	unsigned int flags = cpp_classify_number (parse_in, tok);

	switch (flags & CPP_N_CATEGORY)
	  {
	  case CPP_N_INVALID:
	    /* cpplib has issued an error.  */
	    *value = error_mark_node;
	    errorcount++;
	    break;

	  case CPP_N_INTEGER:
	    *value = interpret_integer (tok, flags);
	    break;

	  case CPP_N_FLOATING:
	    *value = interpret_float (tok, flags);
	    break;

	  default:
	    gcc_unreachable ();
	  }
      }
      break;

    case CPP_ATSIGN:
      /* An @ may give the next token special significance in Objective-C.  */
      if (c_dialect_objc ())
	{
	  location_t atloc = input_location;
	  
	retry_at:
	  tok = cpp_get_token (parse_in);
	  type = tok->type;
	  switch (type)
	    {
	    case CPP_PADDING:
	      goto retry_at;
	      
	    case CPP_STRING:
	    case CPP_WSTRING:
	      type = lex_string (tok, value, true);
	      break;

	    case CPP_NAME:
	      *value = HT_IDENT_TO_GCC_IDENT (HT_NODE (tok->val.node));
	      if (objc_is_reserved_word (*value))
		{
		  type = CPP_AT_NAME;
		  break;
		}
	      /* FALLTHROUGH */

	    default:
	      /* ... or not.  */
	      error ("%Hstray %<@%> in program", &atloc);
	      goto retry_after_at;
	    }
	  break;
	}

      /* FALLTHROUGH */
    case CPP_HASH:
    case CPP_PASTE:
      {
	unsigned char name[4];
	
	*cpp_spell_token (parse_in, tok, name, true) = 0;
	
	error ("stray %qs in program", name);
      }
      
      goto retry;

    case CPP_OTHER:
      {
	cppchar_t c = tok->val.str.text[0];

	if (c == '"' || c == '\'')
	  error ("missing terminating %c character", (int) c);
	else if (ISGRAPH (c))
	  error ("stray %qc in program", (int) c);
	else
	  error ("stray %<\\%o%> in program", (int) c);
      }
      goto retry;

    case CPP_CHAR:
    case CPP_WCHAR:
      *value = lex_charconst (tok);
      break;

    case CPP_STRING:
    case CPP_WSTRING:
      if (!c_lex_return_raw_strings)
	{
	  type = lex_string (tok, value, false);
	  break;
	}
      
      /* FALLTHROUGH */

    case CPP_PRAGMA:
      *value = build_string (tok->val.str.len, (char *) tok->val.str.text);
      break;

      /* These tokens should not be visible outside cpplib.  */
    case CPP_HEADER_NAME:
    case CPP_COMMENT:
    case CPP_MACRO_ARG:
      gcc_unreachable ();

    default:
      *value = NULL_TREE;
      break;
    }

  if (cpp_flags)
    *cpp_flags = tok->flags;

  if (!no_more_pch)
    {
      no_more_pch = true;
      c_common_no_more_pch ();
    }
  
  timevar_pop (TV_CPP);
  
  return type;
}
Beispiel #6
0
enum cpp_ttype
/* APPLE LOCAL CW asm blocks C++ comments 6338079 */
c_lex_with_flags (tree *value, location_t *loc, unsigned char *cpp_flags, int defer)
{
  static bool no_more_pch;
  const cpp_token *tok;
  enum cpp_ttype type;
  unsigned char add_flags = 0;
  /* APPLE LOCAL CW asm blocks C++ */
  const cpp_token *lasttok;
  /* APPLE LOCAL begin CW asm blocks */
  /* Make a local copy of the flag for efficiency, since the compiler can't
     figure that it won't change during a compilation.  */
  int flag_iasm_blocks_local = flag_iasm_blocks;
  if (flag_iasm_blocks_local)
    ++c_lex_depth;
  /* APPLE LOCAL end CW asm blocks */

  timevar_push (TV_CPP);
 retry:
  /* APPLE LOCAL begin CW asm blocks */
  /* If there's a token we saved while returning the special BOL
     token, return it now.  */
  if (iasm_saved_token)
    {
      tok = iasm_saved_token;
      type = tok->type;
      iasm_saved_token = NULL;
      goto bypass;
    }
  /* APPLE LOCAL end CW asm blocks */
  tok = cpp_get_token (parse_in);
  type = tok->type;

  /* APPLE LOCAL begin CW asm blocks */
  /* This test should be as efficient as possible, because it affects
       all lexing with or without CW asm enabled.  */
  if (flag_iasm_blocks_local && iasm_state != iasm_none && c_lex_depth == 1
      && type != CPP_PADDING)
    {
      /* "}" switches us out of our special mode.  */
      if (tok->type == CPP_CLOSE_BRACE && iasm_state >= iasm_decls)
	iasm_state = iasm_none;

      /* This is tricky.  We're only ready to start parsing assembly
	 instructions if we're in the asm block, we're not in the
	 middle of parsing a C decl, and the next token is plausibly
	 the beginning of an asm line.  This works because if we have
	 a "typedef int nop", a nop at the beginning of a line should
	 be taken as an instruction rather than a declaration of type
	 nop.  (Doesn't have to go this way, but it's how CW works.)
	 We're not quite as good as CW yet, because CW knows about the
	 complete list of valid opcodes, and will try to take anything
	 as a decl that is not in the opcode list.  */
      if (iasm_state == iasm_decls
	  && !iasm_in_decl)
	{
	  if (tok->type == CPP_ATSIGN
	      || tok->type == CPP_DOT
	      || (tok->type == CPP_SEMICOLON)
	      || (tok->type == CPP_NAME
		  && (*value = HT_IDENT_TO_GCC_IDENT (HT_NODE (tok->val.node)))
		  && !iasm_typename_or_reserved (*value)))
	    {
	      iasm_state = iasm_asm;
	      inside_iasm_block = true;
	    }
	  else
	    {
	      iasm_in_decl = true;
	    }
	}
    }
 bypass:
  /* APPLE LOCAL end CW asm blocks */

 retry_after_at:
#ifdef USE_MAPPED_LOCATION
  *loc = tok->src_loc;
#else
  *loc = input_location;
#endif
  switch (type)
    {
    case CPP_PADDING:
      goto retry;

    case CPP_NAME:
      *value = HT_IDENT_TO_GCC_IDENT (HT_NODE (tok->val.node));
      /* APPLE LOCAL begin CW asm blocks */
      if (iasm_state >= iasm_decls
	  && flag_ms_asms
	  && strcasecmp (IDENTIFIER_POINTER (HT_IDENT_TO_GCC_IDENT (HT_NODE (tok->val.node))),
			 "offset") == 0)
	{
	  type = CPP_AND;
	  *value = NULL_TREE;
	  break;
	}
      /* APPLE LOCAL end CW asm blocks */
      break;

    case CPP_NUMBER:
      {
	/* APPLE LOCAL CW asm blocks C++ comments 6338079 */
	unsigned int flags = cpp_classify_number (parse_in, tok, defer);

	switch (flags & CPP_N_CATEGORY)
	  {
	  case CPP_N_INVALID:
	    /* APPLE LOCAL begin CW asm blocks C++ comments 6338079 */
	    if (flags & CPP_N_DEFER)
	      {
		add_flags = ERROR_DEFERRED;
		*value = error_mark_node;
		break;
	      }
	    /* APPLE LOCAL end CW asm blocks C++ comments 6338079 */

	    /* cpplib has issued an error.  */
	    *value = error_mark_node;
	    errorcount++;
	    break;

	  case CPP_N_INTEGER:
	    /* C++ uses '0' to mark virtual functions as pure.
	       Set PURE_ZERO to pass this information to the C++ parser.  */
	    if (tok->val.str.len == 1 && *tok->val.str.text == '0')
	      add_flags = PURE_ZERO;
	    *value = interpret_integer (tok, flags);
	    break;

	  case CPP_N_FLOATING:
	    *value = interpret_float (tok, flags);
	    break;

	  default:
	    gcc_unreachable ();
	  }
      }
      break;

    case CPP_ATSIGN:
      /* APPLE LOCAL begin CW asm blocks */
      if (iasm_state >= iasm_decls)
	{
	  /* Return the @-sign verbatim.  */
	  *value = NULL_TREE;
	  break;
	}
      lasttok = tok;
      /* APPLE LOCAL end CW asm blocks */

      /* An @ may give the next token special significance in Objective-C.  */
      if (c_dialect_objc ())
	{
	  location_t atloc = input_location;

	  /* APPLE LOCAL CW asm blocks */
	  ++parse_in->keep_tokens;
	retry_at:
	  tok = cpp_get_token (parse_in);
	  type = tok->type;
	  switch (type)
	    {
	    case CPP_PADDING:
	      goto retry_at;

	    case CPP_STRING:
	    case CPP_WSTRING:
	      type = lex_string (tok, value, true);
	      break;

	    case CPP_NAME:
	      *value = HT_IDENT_TO_GCC_IDENT (HT_NODE (tok->val.node));
	      if (objc_is_reserved_word (*value))
		{
		  type = CPP_AT_NAME;
		  break;
		}
	      /* FALLTHROUGH */

	    default:
	      /* APPLE LOCAL begin CW asm blocks C++ */
	      if (flag_iasm_blocks_local)
                {
                  /* This is necessary for C++, as we don't have the tight
                     integration between the lexer and the parser... */
		  _cpp_backup_tokens (parse_in, 1);
                  /* Return the @-sign verbatim.  */
                  *value = NULL;
                  tok = lasttok;
		  type = tok->type;
                  break;
                }
	      /* APPLE LOCAL end CW asm blocks C++ */

	      /* ... or not.  */
	      error ("%Hstray %<@%> in program", &atloc);
	      goto retry_after_at;
	    }
	  /* APPLE LOCAL CW asm blocks */
	  --parse_in->keep_tokens;
	  break;
	}
	/* APPLE LOCAL begin CW asm blocks C++ */
	if (flag_iasm_blocks_local)
	  {
	    /* This is necessary for C++, as we don't have the tight
	       integration between the lexer and the parser... */
	    /* Return the @-sign verbatim.  */
	    *value = NULL;
	    break;
	  }
       /* APPLE LOCAL end CW asm blocks C++ */
	      
      /* FALLTHROUGH */
    case CPP_HASH:
    case CPP_PASTE:
      /* APPLE LOCAL begin CW asm blocks C++ comments 4248139 */
      /* Because we don't recognize inline asm commments during
	 lexing, we have to pass this back to the parser to error out
	 with or eat as a comment as appropriate.  */
      if (flag_iasm_blocks_local)
	{
	  *value = NULL;
	  break;
	}
      /* APPLE LOCAL end CW asm blocks C++ comments 4248139 */
      {
	unsigned char name[4];

	*cpp_spell_token (parse_in, tok, name, true) = 0;

	error ("stray %qs in program", name);
      }

      goto retry;

    case CPP_OTHER:
      {
	cppchar_t c = tok->val.str.text[0];

	/* APPLE LOCAL begin CW asm blocks C++ comments 4248139 */
	/* Because we don't recognize inline asm commments during
	   lexing, we have to pass this back to the parser to error
	   out with or eat as a comment as appropriate.  */
	if (defer && flag_iasm_blocks_local)
	  {
	    *value = build_int_cst_wide (char_type_node, c, 0);
	    break;
	  }
	/* APPLE LOCAL end CW asm blocks C++ comments 4248139 */
	if (c == '"' || c == '\'')
	  error ("missing terminating %c character", (int) c);
	else if (ISGRAPH (c))
	  error ("stray %qc in program", (int) c);
	else
	  error ("stray %<\\%o%> in program", (int) c);
      }
      goto retry;

    case CPP_CHAR:
    case CPP_WCHAR:
      *value = lex_charconst (tok);
      break;

    case CPP_STRING:
    case CPP_WSTRING:
      if (!c_lex_return_raw_strings)
	{
	  type = lex_string (tok, value, false);
	  break;
	}
      *value = build_string (tok->val.str.len, (char *) tok->val.str.text);
      break;
      
    case CPP_PRAGMA:
      *value = build_int_cst (NULL, tok->val.pragma);
      break;

    /* APPLE LOCAL begin CW asm blocks */
    case CPP_MULT:
      if (inside_iasm_block)
        {
	  /* Check and replace use of '*' with '.' if '*' is followed by '-'
	     or '+'. This is to allow "b *+8" which is disallwed by darwin's
	     assembler but nevertheless is needed to be compatible with CW tools. */
  	  lasttok = tok;
	  ++parse_in->keep_tokens;
	  do
              tok = cpp_get_token (parse_in);
          while (tok->type == CPP_PADDING);
	  _cpp_backup_tokens (parse_in, 1);
	  --parse_in->keep_tokens;
	  if (tok->type == CPP_PLUS || tok->type == CPP_MINUS)
	      type = CPP_DOT;
	  tok = lasttok;
        }
      *value = NULL_TREE;
      break;
    /* APPLE LOCAL end CW asm blocks */
    /* APPLE LOCAL begin 4137741 */
    /* For CPP_BINCL and CPP_EINCL tokens, we shall need to propagate
       line number information; the location field shall already include
       the desired file name.  */
    case CPP_BINCL:
    case CPP_EINCL:
      *value = build_int_cst (integer_type_node, (HOST_WIDE_INT) tok->src_loc);
      break;

    /* APPLE LOCAL end 4137741 */
      /* These tokens should not be visible outside cpplib.  */
    case CPP_HEADER_NAME:
    case CPP_COMMENT:
    case CPP_MACRO_ARG:
      gcc_unreachable ();

    default:
      *value = NULL_TREE;
      break;
    }

  if (cpp_flags)
    *cpp_flags = tok->flags | add_flags;

  /* APPLE LOCAL begin CW asm blocks */
  if (flag_iasm_blocks_local)
    --c_lex_depth;
  /* APPLE LOCAL end CW asm blocks */

    /* APPLE LOCAL begin 4137741 */
  if (!no_more_pch
      && type != CPP_BINCL
      && type != CPP_EINCL)
    /* APPLE LOCAL end 4137741 */
    {
      no_more_pch = true;
      c_common_no_more_pch ();
    }

  timevar_pop (TV_CPP);

  return type;
}
Beispiel #7
0
/* Match pattern "p" against "text" */
static int dowild(const uchar *p, const uchar *text, unsigned int flags)
{
	uchar p_ch;
	const uchar *pattern = p;

	int pattern_needs_leading_period =
		(flags & WM_PERIOD) && (text[0] == '.');

	for ( ; (p_ch = *p) != '\0'; text++, p++) {
		int matched, match_slash, negated;
		uchar t_ch, prev_ch;
		if ((t_ch = *text) == '\0' && p_ch != '*')
			return WM_ABORT_ALL;
		if ((flags & WM_CASEFOLD) && ISUPPER(t_ch))
			t_ch = tolower(t_ch);
		if ((flags & WM_CASEFOLD) && ISUPPER(p_ch))
			p_ch = tolower(p_ch);
		switch (p_ch) {
		case '\\':
			if (!(flags & WM_NOESCAPE))
				/* Literal match with following character.  Note that the test
				 * in "default" handles the p[1] == '\0' failure case. */
				p_ch = *++p;
			/* FALLTHROUGH */
		default:
			if (t_ch != p_ch)
				return WM_NOMATCH;
			if ((flags & WM_PERIOD) &&
			    (flags & WM_PATHNAME) && t_ch == '/' &&
			    text[1] == '.' && p[1] != '.')
				return WM_NOMATCH;
			/* If we needed a leading period, we've matched it. */
			pattern_needs_leading_period = 0;
			continue;
		case '?':
			/* Match anything but '/'. */
			if ((flags & WM_PATHNAME) && t_ch == '/')
				return WM_NOMATCH;
			continue;
		case '*':
			if (*++p == '*') {
				const uchar *prev_p = p - 2;
				while (*++p == '*') {}
				if (!(flags & WM_PATHNAME))
					/* without WM_PATHNAME, '*' == '**' */
					match_slash = 1;
				else if ((prev_p < pattern || *prev_p == '/') &&
				    (*p == '\0' || *p == '/' ||
				     (p[0] == '\\' && p[1] == '/'))) {
					/*
					 * Assuming we already match 'foo/' and are at
					 * <star star slash>, just assume it matches
					 * nothing and go ahead match the rest of the
					 * pattern with the remaining string. This
					 * helps make foo/<*><*>/bar (<> because
					 * otherwise it breaks C comment syntax) match
					 * both foo/bar and foo/a/bar.
					 */
					if (p[0] == '/' &&
					    dowild(p + 1, text, flags) == WM_MATCH)
						return WM_MATCH;
					match_slash = 1;
				} else
					return WM_ABORT_MALFORMED;
			} else
				/* without WM_PATHNAME, '*' == '**' */
				match_slash = flags & WM_PATHNAME ? 0 : 1;
			if (*p == '\0') {
				/* If we needed a leading period in the pattern but only
				 * found stars, we didn't match. */
				if (pattern_needs_leading_period)
					return WM_NOMATCH;
				/* Trailing "**" matches everything.  Trailing "*" matches
				 * only if there are no more slash characters. */
				if (!match_slash) {
					if (strchr((char*)text, '/') != NULL)
						return WM_NOMATCH;
					if ((flags & WM_PERIOD) &&
					    strstr((const char*)text, "/.") != NULL)
						return WM_NOMATCH;
				}
				return WM_MATCH;
			} else if (!match_slash && *p == '/') {
				/*
				 * _one_ asterisk followed by a slash
				 * with WM_PATHNAME matches the next
				 * directory
				 */
				const char *slash = strchr((char*)text, '/');
				if (!slash)
					return WM_NOMATCH;
				text = (const uchar*)slash;
				/* the slash is consumed by the top-level for loop */
				break;
			}
			while (1) {
				if (t_ch == '\0')
					break;
				/*
				 * Try to advance faster when an asterisk is
				 * followed by a literal. We know in this case
				 * that the the string before the literal
				 * must belong to "*".
				 * If match_slash is false, do not look past
				 * the first slash as it cannot belong to '*'.
				 */
				if (!is_glob_special(*p)) {
					p_ch = *p;
					if ((flags & WM_CASEFOLD) && ISUPPER(p_ch))
						p_ch = tolower(p_ch);
					while ((t_ch = *text) != '\0' &&
					       (match_slash || t_ch != '/')) {
						if ((flags & WM_CASEFOLD) && ISUPPER(t_ch))
							t_ch = tolower(t_ch);
						if (t_ch == p_ch)
							break;
						text++;
					}
					if (t_ch != p_ch)
						return WM_NOMATCH;
				}
				if (pattern_needs_leading_period)
					return WM_NOMATCH;
				if ((matched = dowild(p, text, flags)) != WM_NOMATCH) {
					if (!match_slash || matched != WM_ABORT_TO_STARSTAR)
						return matched;
				} else if (!match_slash && t_ch == '/')
					return WM_ABORT_TO_STARSTAR;
				else if ((flags & WM_PERIOD) &&
					 (flags & WM_PATHNAME) && t_ch == '/' &&
					 text[1] == '.' && p[1] != '.')
					return WM_NOMATCH;
				t_ch = *++text;
			}
			return WM_ABORT_ALL;
		case '[':
			p_ch = *++p;
#ifdef NEGATE_CLASS2
			if (p_ch == NEGATE_CLASS2)
				p_ch = NEGATE_CLASS;
#endif
			/* Assign literal 1/0 because of "matched" comparison. */
			negated = p_ch == NEGATE_CLASS ? 1 : 0;
			if (negated) {
				/* Inverted character class. */
				p_ch = *++p;
			}
			prev_ch = 0;
			matched = 0;
			do {
				if (!p_ch)
					return WM_ABORT_ALL;
				if (p_ch == '\\') {
					p_ch = *++p;
					if (!p_ch)
						return WM_ABORT_ALL;
					if (t_ch == p_ch)
						matched = 1;
				} else if (p_ch == '-' && prev_ch && p[1] && p[1] != ']') {
					p_ch = *++p;
					if (p_ch == '\\') {
						p_ch = *++p;
						if (!p_ch)
							return WM_ABORT_ALL;
					}
					if (t_ch <= p_ch && t_ch >= prev_ch)
						matched = 1;
					else if ((flags & WM_CASEFOLD) && ISLOWER(t_ch)) {
						uchar t_ch_upper = toupper(t_ch);
						if (t_ch_upper <= p_ch && t_ch_upper >= prev_ch)
							matched = 1;
					}
					p_ch = 0; /* This makes "prev_ch" get set to 0. */
				} else if (p_ch == '[' && p[1] == ':') {
					const uchar *s;
					int i;
					for (s = p += 2; (p_ch = *p) && p_ch != ']'; p++) {} /*SHARED ITERATOR*/
					if (!p_ch)
						return WM_ABORT_ALL;
					i = p - s - 1;
					if (i < 0 || p[-1] != ':') {
						/* Didn't find ":]", so treat like a normal set. */
						p = s - 2;
						p_ch = '[';
						if (t_ch == p_ch)
							matched = 1;
						continue;
					}
					if (CC_EQ(s,i, "alnum")) {
						if (ISALNUM(t_ch))
							matched = 1;
					} else if (CC_EQ(s,i, "alpha")) {
						if (ISALPHA(t_ch))
							matched = 1;
					} else if (CC_EQ(s,i, "blank")) {
						if (ISBLANK(t_ch))
							matched = 1;
					} else if (CC_EQ(s,i, "cntrl")) {
						if (ISCNTRL(t_ch))
							matched = 1;
					} else if (CC_EQ(s,i, "digit")) {
						if (ISDIGIT(t_ch))
							matched = 1;
					} else if (CC_EQ(s,i, "graph")) {
						if (ISGRAPH(t_ch))
							matched = 1;
					} else if (CC_EQ(s,i, "lower")) {
						if (ISLOWER(t_ch))
							matched = 1;
					} else if (CC_EQ(s,i, "print")) {
						if (ISPRINT(t_ch))
							matched = 1;
					} else if (CC_EQ(s,i, "punct")) {
						if (ISPUNCT(t_ch))
							matched = 1;
					} else if (CC_EQ(s,i, "space")) {
						if (ISSPACE(t_ch))
							matched = 1;
					} else if (CC_EQ(s,i, "upper")) {
						if (ISUPPER(t_ch))
							matched = 1;
						else if ((flags & WM_CASEFOLD) && ISLOWER(t_ch))
							matched = 1;
					} else if (CC_EQ(s,i, "xdigit")) {
						if (ISXDIGIT(t_ch))
							matched = 1;
					} else /* malformed [:class:] string */
						return WM_ABORT_ALL;
					p_ch = 0; /* This makes "prev_ch" get set to 0. */
				} else if (t_ch == p_ch)
					matched = 1;
			} while (prev_ch = p_ch, (p_ch = *++p) != ']');
			if (matched == negated ||
			    ((flags & WM_PATHNAME) && t_ch == '/'))
				return WM_NOMATCH;
			if ((flags & WM_PATHNAME) && t_ch == '/' &&
			    (flags & WM_PERIOD) && (text[1] == '.') && (p[1] != '.'))
				return WM_NOMATCH;
			continue;
		}
	}

	return *text ? WM_NOMATCH : WM_MATCH;
}
Beispiel #8
0
/* Match pattern "p" against the a virtually-joined string consisting
 * of "text" and any strings in array "a". */
static int dowild(const uchar *p, const uchar *text, const uchar*const *a)
{
	uchar p_ch;

#ifdef WILD_TEST_ITERATIONS
	wildmatch_iteration_count++;
#endif

	for ( ; (p_ch = *p) != '\0'; text++, p++) {
		int matched, special;
		uchar t_ch, prev_ch;
		while ((t_ch = *text) == '\0') {
			if (*a == NULL) {
				if (p_ch != '*')
					return ABORT_ALL;
				break;
			}
			text = *a++;
		}
		if (force_lower_case && ISUPPER(t_ch))
			t_ch = tolower(t_ch);
		switch (p_ch) {
			case '\\':
				/* Literal match with following character.  Note that the test
				 * in "default" handles the p[1] == '\0' failure case. */
				p_ch = *++p;
				/* FALLTHROUGH */
			default:
				if (t_ch != p_ch)
					return FALSE;
				continue;
			case '?':
				/* Match anything but '/'. */
				if (t_ch == '/')
					return FALSE;
				continue;
			case '*':
				if (*++p == '*') {
					while (*++p == '*') {}
					special = TRUE;
				} else
					special = FALSE;
				if (*p == '\0') {
					/* Trailing "**" matches everything.  Trailing "*" matches
					 * only if there are no more slash characters. */
					if (!special) {
						do {
							if (strchr((char*)text, '/') != NULL)
								return FALSE;
						} while ((text = *a++) != NULL);
					}
					return TRUE;
				}
				while (1) {
					if (t_ch == '\0') {
						if ((text = *a++) == NULL)
							break;
						t_ch = *text;
						continue;
					}
					if ((matched = dowild(p, text, a)) != FALSE) {
						if (!special || matched != ABORT_TO_STARSTAR)
							return matched;
					} else if (!special && t_ch == '/')
						return ABORT_TO_STARSTAR;
					t_ch = *++text;
				}
				return ABORT_ALL;
			case '[':
				p_ch = *++p;
#ifdef NEGATE_CLASS2
				if (p_ch == NEGATE_CLASS2)
					p_ch = NEGATE_CLASS;
#endif
				/* Assign literal TRUE/FALSE because of "matched" comparison. */
				special = p_ch == NEGATE_CLASS? TRUE : FALSE;
				if (special) {
					/* Inverted character class. */
					p_ch = *++p;
				}
				prev_ch = 0;
				matched = FALSE;
				do {
					if (!p_ch)
						return ABORT_ALL;
					if (p_ch == '\\') {
						p_ch = *++p;
						if (!p_ch)
							return ABORT_ALL;
						if (t_ch == p_ch)
							matched = TRUE;
					} else if (p_ch == '-' && prev_ch && p[1] && p[1] != ']') {
						p_ch = *++p;
						if (p_ch == '\\') {
							p_ch = *++p;
							if (!p_ch)
								return ABORT_ALL;
						}
						if (t_ch <= p_ch && t_ch >= prev_ch)
							matched = TRUE;
						p_ch = 0; /* This makes "prev_ch" get set to 0. */
					} else if (p_ch == '[' && p[1] == ':') {
						const uchar *s;
						int i;
						for (s = p += 2; (p_ch = *p) && p_ch != ']'; p++) {}
						if (!p_ch)
							return ABORT_ALL;
						i = p - s - 1;
						if (i < 0 || p[-1] != ':') {
							/* Didn't find ":]", so treat like a normal set. */
							p = s - 2;
							p_ch = '[';
							if (t_ch == p_ch)
								matched = TRUE;
							continue;
						}
						if (CC_EQ(s,i, "alnum")) {
							if (ISALNUM(t_ch))
								matched = TRUE;
						} else if (CC_EQ(s,i, "alpha")) {
							if (ISALPHA(t_ch))
								matched = TRUE;
						} else if (CC_EQ(s,i, "blank")) {
							if (ISBLANK(t_ch))
								matched = TRUE;
						} else if (CC_EQ(s,i, "cntrl")) {
							if (ISCNTRL(t_ch))
								matched = TRUE;
						} else if (CC_EQ(s,i, "digit")) {
							if (ISDIGIT(t_ch))
								matched = TRUE;
						} else if (CC_EQ(s,i, "graph")) {
							if (ISGRAPH(t_ch))
								matched = TRUE;
						} else if (CC_EQ(s,i, "lower")) {
							if (ISLOWER(t_ch))
								matched = TRUE;
						} else if (CC_EQ(s,i, "print")) {
							if (ISPRINT(t_ch))
								matched = TRUE;
						} else if (CC_EQ(s,i, "punct")) {
							if (ISPUNCT(t_ch))
								matched = TRUE;
						} else if (CC_EQ(s,i, "space")) {
							if (ISSPACE(t_ch))
								matched = TRUE;
						} else if (CC_EQ(s,i, "upper")) {
							if (ISUPPER(t_ch))
								matched = TRUE;
						} else if (CC_EQ(s,i, "xdigit")) {
							if (ISXDIGIT(t_ch))
								matched = TRUE;
						} else /* malformed [:class:] string */
							return ABORT_ALL;
						p_ch = 0; /* This makes "prev_ch" get set to 0. */
					} else if (t_ch == p_ch)
						matched = TRUE;
				} while (prev_ch = p_ch, (p_ch = *++p) != ']');
				if (matched == special || t_ch == '/')
					return FALSE;
				continue;
		}
	}

	do {
		if (*text)
			return FALSE;
	} while ((text = *a++) != NULL);

	return TRUE;
}
Beispiel #9
0
static int
internal_function
FCT (const CHAR *pattern, const CHAR *string, const CHAR *string_end,
     bool no_leading_period, int flags)
{
  register const CHAR *p = pattern, *n = string;
  register UCHAR c;
#ifdef _LIBC
# if WIDE_CHAR_VERSION
  const char *collseq = (const char *)
    _NL_CURRENT(LC_COLLATE, _NL_COLLATE_COLLSEQWC);
# else
  const UCHAR *collseq = (const UCHAR *)
    _NL_CURRENT(LC_COLLATE, _NL_COLLATE_COLLSEQMB);
# endif
#endif

  while ((c = *p++) != L_('\0'))
    {
      bool new_no_leading_period = false;
      c = FOLD (c);

      switch (c)
	{
	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;
	    }

	  if (n == string_end)
	    return FNM_NOMATCH;
	  else if (*n == L_('/') && (flags & FNM_FILE_NAME))
	    return FNM_NOMATCH;
	  else if (*n == L_('.') && no_leading_period)
	    return FNM_NOMATCH;
	  break;

	case L_('\\'):
	  if (!(flags & FNM_NOESCAPE))
	    {
	      c = *p++;
	      if (c == L_('\0'))
		/* Trailing \ loses.  */
		return FNM_NOMATCH;
	      c = FOLD (c);
	    }
	  if (n == string_end || FOLD ((UCHAR) *n) != c)
	    return FNM_NOMATCH;
	  break;

	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;
	    }

	  if (n != string_end && *n == L_('.') && no_leading_period)
	    return FNM_NOMATCH;

	  for (c = *p++; c == L_('?') || c == L_('*'); c = *p++)
	    {
	      if (*p == L_('(') && (flags & FNM_EXTMATCH) != 0)
		{
		  const CHAR *endp = END (p);
		  if (endp != p)
		    {
		      /* This is a pattern.  Skip over it.  */
		      p = endp;
		      continue;
		    }
		}

	      if (c == L_('?'))
		{
		  /* A ? needs to match one character.  */
		  if (n == string_end)
		    /* There isn't another character; no match.  */
		    return FNM_NOMATCH;
		  else if (*n == L_('/')
			   && __builtin_expect (flags & FNM_FILE_NAME, 0))
		    /* A slash does not match a wildcard under
		       FNM_FILE_NAME.  */
		    return FNM_NOMATCH;
		  else
		    /* One character of the string is consumed in matching
		       this ? wildcard, so *??? won't match if there are
		       less than three characters.  */
		    ++n;
		}
	    }

	  if (c == L_('\0'))
	    /* The wildcard(s) is/are the last element of the pattern.
	       If the name is a file name and contains another slash
	       this means it cannot match, unless the FNM_LEADING_DIR
	       flag is set.  */
	    {
	      int result = (flags & FNM_FILE_NAME) == 0 ? 0 : FNM_NOMATCH;

	      if (flags & FNM_FILE_NAME)
		{
		  if (flags & FNM_LEADING_DIR)
		    result = 0;
		  else
		    {
		      if (MEMCHR (n, L_('/'), string_end - n) == NULL)
			result = 0;
		    }
		}

	      return result;
	    }
	  else
	    {
	      const CHAR *endp;

	      endp = MEMCHR (n, (flags & FNM_FILE_NAME) ? L_('/') : L_('\0'),
			     string_end - n);
	      if (endp == NULL)
		endp = string_end;

	      if (c == L_('[')
		  || (__builtin_expect (flags & FNM_EXTMATCH, 0) != 0
		      && (c == L_('@') || c == L_('+') || c == L_('!'))
		      && *p == L_('(')))
		{
		  int flags2 = ((flags & FNM_FILE_NAME)
				? flags : (flags & ~FNM_PERIOD));
		  bool no_leading_period2 = no_leading_period;

		  for (--p; n < endp; ++n, no_leading_period2 = false)
		    if (FCT (p, n, string_end, no_leading_period2, flags2)
			== 0)
		      return 0;
		}
	      else if (c == L_('/') && (flags & FNM_FILE_NAME))
		{
		  while (n < string_end && *n != L_('/'))
		    ++n;
		  if (n < string_end && *n == L_('/')
		      && (FCT (p, n + 1, string_end, flags & FNM_PERIOD, flags)
			  == 0))
		    return 0;
		}
	      else
		{
		  int flags2 = ((flags & FNM_FILE_NAME)
				? flags : (flags & ~FNM_PERIOD));
		  int no_leading_period2 = no_leading_period;

		  if (c == L_('\\') && !(flags & FNM_NOESCAPE))
		    c = *p;
		  c = FOLD (c);
		  for (--p; n < endp; ++n, no_leading_period2 = false)
		    if (FOLD ((UCHAR) *n) == c
			&& (FCT (p, n, string_end, no_leading_period2, flags2)
			    == 0))
		      return 0;
		}
	    }

	  /* If we come here no match is possible with the wildcard.  */
	  return FNM_NOMATCH;

	case L_('['):
	  {
	    /* Nonzero if the sense of the character class is inverted.  */
	    register bool not;
	    CHAR cold;
	    UCHAR fn;

	    if (posixly_correct == 0)
	      posixly_correct = getenv ("POSIXLY_CORRECT") != NULL ? 1 : -1;

	    if (n == string_end)
	      return FNM_NOMATCH;

	    if (*n == L_('.') && no_leading_period)
	      return FNM_NOMATCH;

	    if (*n == L_('/') && (flags & FNM_FILE_NAME))
	      /* `/' cannot be matched.  */
	      return FNM_NOMATCH;

	    not = (*p == L_('!') || (posixly_correct < 0 && *p == L_('^')));
	    if (not)
	      ++p;

	    fn = FOLD ((UCHAR) *n);

	    c = *p++;
	    for (;;)
	      {
		if (!(flags & FNM_NOESCAPE) && c == L_('\\'))
		  {
		    if (*p == L_('\0'))
		      return FNM_NOMATCH;
		    c = FOLD ((UCHAR) *p);
		    ++p;

		    if (c == fn)
		      goto matched;
		  }
		else if (c == L_('[') && *p == L_(':'))
		  {
		    /* Leave room for the null.  */
		    CHAR str[CHAR_CLASS_MAX_LENGTH + 1];
		    size_t c1 = 0;
#if defined _LIBC || WIDE_CHAR_SUPPORT
		    wctype_t wt;
#endif
		    const CHAR *startp = p;

		    for (;;)
		      {
			if (c1 == CHAR_CLASS_MAX_LENGTH)
			  /* The name is too long and therefore the pattern
			     is ill-formed.  */
			  return FNM_NOMATCH;

			c = *++p;
			if (c == L_(':') && p[1] == L_(']'))
			  {
			    p += 2;
			    break;
			  }
			if (c < L_('a') || c >= L_('z'))
			  {
			    /* This cannot possibly be a character class name.
			       Match it as a normal range.  */
			    p = startp;
			    c = L_('[');
			    goto normal_bracket;
			  }
			str[c1++] = c;
		      }
		    str[c1] = L_('\0');

#if defined _LIBC || WIDE_CHAR_SUPPORT
		    wt = IS_CHAR_CLASS (str);
		    if (wt == 0)
		      /* Invalid character class name.  */
		      return FNM_NOMATCH;

# if defined _LIBC && ! WIDE_CHAR_VERSION
		    /* The following code is glibc specific but does
		       there a good job in speeding up the code since
		       we can avoid the btowc() call.  */
		    if (_ISCTYPE ((UCHAR) *n, wt))
		      goto matched;
# else
		    if (ISWCTYPE (BTOWC ((UCHAR) *n), wt))
		      goto matched;
# endif
#else
		    if ((STREQ (str, L_("alnum")) && ISALNUM ((UCHAR) *n))
			|| (STREQ (str, L_("alpha")) && ISALPHA ((UCHAR) *n))
			|| (STREQ (str, L_("blank")) && ISBLANK ((UCHAR) *n))
			|| (STREQ (str, L_("cntrl")) && ISCNTRL ((UCHAR) *n))
			|| (STREQ (str, L_("digit")) && ISDIGIT ((UCHAR) *n))
			|| (STREQ (str, L_("graph")) && ISGRAPH ((UCHAR) *n))
			|| (STREQ (str, L_("lower")) && ISLOWER ((UCHAR) *n))
			|| (STREQ (str, L_("print")) && ISPRINT ((UCHAR) *n))
			|| (STREQ (str, L_("punct")) && ISPUNCT ((UCHAR) *n))
			|| (STREQ (str, L_("space")) && ISSPACE ((UCHAR) *n))
			|| (STREQ (str, L_("upper")) && ISUPPER ((UCHAR) *n))
			|| (STREQ (str, L_("xdigit")) && ISXDIGIT ((UCHAR) *n)))
		      goto matched;
#endif
		    c = *p++;
		  }
#ifdef _LIBC
		else if (c == L_('[') && *p == L_('='))
		  {
		    UCHAR str[1];
		    uint32_t nrules =
		      _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES);
		    const CHAR *startp = p;

		    c = *++p;
		    if (c == L_('\0'))
		      {
			p = startp;
			c = L_('[');
			goto normal_bracket;
		      }
		    str[0] = c;

		    c = *++p;
		    if (c != L_('=') || p[1] != L_(']'))
		      {
			p = startp;
			c = L_('[');
			goto normal_bracket;
		      }
		    p += 2;

		    if (nrules == 0)
		      {
			if ((UCHAR) *n == str[0])
			  goto matched;
		      }
		    else
		      {
			const int32_t *table;
# if WIDE_CHAR_VERSION
			const int32_t *weights;
			const int32_t *extra;
# else
			const unsigned char *weights;
			const unsigned char *extra;
# endif
			const int32_t *indirect;
			int32_t idx;
			const UCHAR *cp = (const UCHAR *) str;

			/* This #include defines a local function!  */
# if WIDE_CHAR_VERSION
#  include <locale/weightwc.h>
# else
#  include <locale/weight.h>
# endif

# if WIDE_CHAR_VERSION
			table = (const int32_t *)
			  _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEWC);
			weights = (const int32_t *)
			  _NL_CURRENT (LC_COLLATE, _NL_COLLATE_WEIGHTWC);
			extra = (const int32_t *)
			  _NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAWC);
			indirect = (const int32_t *)
			  _NL_CURRENT (LC_COLLATE, _NL_COLLATE_INDIRECTWC);
# else
			table = (const int32_t *)
			  _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB);
			weights = (const unsigned char *)
			  _NL_CURRENT (LC_COLLATE, _NL_COLLATE_WEIGHTMB);
			extra = (const unsigned char *)
			  _NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAMB);
			indirect = (const int32_t *)
			  _NL_CURRENT (LC_COLLATE, _NL_COLLATE_INDIRECTMB);
# endif

			idx = findidx (&cp);
			if (idx != 0)
			  {
			    /* We found a table entry.  Now see whether the
			       character we are currently at has the same
			       equivalance class value.  */
			    int len = weights[idx];
			    int32_t idx2;
			    const UCHAR *np = (const UCHAR *) n;

			    idx2 = findidx (&np);
			    if (idx2 != 0 && len == weights[idx2])
			      {
				int cnt = 0;

				while (cnt < len
				       && (weights[idx + 1 + cnt]
					   == weights[idx2 + 1 + cnt]))
				  ++cnt;

				if (cnt == len)
				  goto matched;
			      }
			  }
		      }

		    c = *p++;
		  }
#endif
		else if (c == L_('\0'))
		  /* [ (unterminated) loses.  */
		  return FNM_NOMATCH;
		else
		  {
		    bool is_range = false;

#ifdef _LIBC
		    bool is_seqval = false;

		    if (c == 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;
			  }

			/* We have to handling the symbols differently in
			   ranges since then the collation sequence is
			   important.  */
			is_range = *p == L_('-') && p[1] != L_('\0');

			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;

			    if (!is_range && *n == startp[1])
			      goto matched;

			    cold = startp[1];
			    c = *p++;
			  }
			else
			  {
			    int32_t table_size;
			    const int32_t *symb_table;
# ifdef WIDE_CHAR_VERSION
			    char str[c1];
			    size_t 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) & ~3;

				wextra = (int32_t *) &extra[idx + 4];
# endif

				if (! is_range)
				  {
# ifdef WIDE_CHAR_VERSION
				    for (c1 = 0;
					 (int32_t) c1 < wextra[idx];
					 ++c1)
				      if (n[c1] != wextra[1 + c1])
					break;

				    if ((int32_t) c1 == wextra[idx])
				      goto matched;
# else
				    for (c1 = 0; c1 < extra[idx]; ++c1)
				      if (n[c1] != extra[1 + c1])
					break;

				    if (c1 == extra[idx])
				      goto matched;
# endif
				  }

				/* Get the collation sequence value.  */
				is_seqval = true;
# ifdef WIDE_CHAR_VERSION
				cold = wextra[1 + wextra[idx]];
# else
				/* Adjust for the alignment.  */
				idx += 1 + extra[idx];
				idx = (idx + 3) & ~4;
				cold = *((int32_t *) &extra[idx]);
# endif

				c = *p++;
			      }
			    else if (c1 == 1)
			      {
				/* No valid character.  Match it as a
				   single byte.  */
				if (!is_range && *n == str[0])
				  goto matched;

				cold = str[0];
				c = *p++;
			      }
			    else
			      return FNM_NOMATCH;
			  }
		      }
		    else
# undef str
#endif
		      {
			c = FOLD (c);
		      normal_bracket:

			/* We have to handling the symbols differently in
			   ranges since then the collation sequence is
			   important.  */
			is_range = (*p == L_('-') && p[1] != L_('\0')
				    && p[1] != L_(']'));

			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 = false;
			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];
				size_t 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 = true;
# 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 != (UCHAR) *n)
		return FNM_NOMATCH;

	      new_no_leading_period = true;
	      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;
}
Beispiel #10
0
static int
internal_function
internal_fnmatch (const char *pattern, const char *string, int no_leading_period, int flags)




{
  register const char *p = pattern, *n = string;
  register unsigned char c;

/* Note that this evaluates C many times.  */
# ifdef _LIBC
#  define FOLD(c) ((flags & FNM_CASEFOLD) ? tolower (c) : (c))
# else
#  define FOLD(c) ((flags & FNM_CASEFOLD) && ISUPPER (c) ? tolower (c) : (c))
# endif

  while ((c = *p++) != '\0')
    {
      c = FOLD (c);

      switch (c)
	{
	case '?':
	  if (*n == '\0')
	    return FNM_NOMATCH;
	  else if (*n == '/' && (flags & FNM_FILE_NAME))
	    return FNM_NOMATCH;
	  else if (*n == '.' && no_leading_period
		   && (n == string
		       || (n[-1] == '/' && (flags & FNM_FILE_NAME))))
	    return FNM_NOMATCH;
	  break;

	case '\\':
	  if (!(flags & FNM_NOESCAPE))
	    {
	      c = *p++;
	      if (c == '\0')
		/* Trailing \ loses.  */
		return FNM_NOMATCH;
	      c = FOLD (c);
	    }
	  if (FOLD ((unsigned char) *n) != c)
	    return FNM_NOMATCH;
	  break;

	case '*':
	  if (*n == '.' && no_leading_period
	      && (n == string
		  || (n[-1] == '/' && (flags & FNM_FILE_NAME))))
	    return FNM_NOMATCH;

	  for (c = *p++; c == '?' || c == '*'; c = *p++)
	    {
	      if (*n == '/' && (flags & FNM_FILE_NAME))
		/* A slash does not match a wildcard under FNM_FILE_NAME.  */
		return FNM_NOMATCH;
	      else if (c == '?')
		{
		  /* A ? needs to match one character.  */
		  if (*n == '\0')
		    /* There isn't another character; no match.  */
		    return FNM_NOMATCH;
		  else
		    /* One character of the string is consumed in matching
		       this ? wildcard, so *??? won't match if there are
		       less than three characters.  */
		    ++n;
		}
	    }

	  if (c == '\0')
	    /* The wildcard(s) is/are the last element of the pattern.
	       If the name is a file name and contains another slash
	       this does mean it cannot match.  */
	    return ((flags & FNM_FILE_NAME) && strchr (n, '/') != NULL
		    ? FNM_NOMATCH : 0);
	  else
	    {
	      const char *endp;

	      endp = __strchrnul (n, (flags & FNM_FILE_NAME) ? '/' : '\0');

	      if (c == '[')
		{
		  int flags2 = ((flags & FNM_FILE_NAME)
				? flags : (flags & ~FNM_PERIOD));

		  for (--p; n < endp; ++n)
		    if (internal_fnmatch (p, n,
					  (no_leading_period
					   && (n == string
					       || (n[-1] == '/'
						   && (flags
						       & FNM_FILE_NAME)))),
					  flags2)
			== 0)
		      return 0;
		}
	      else if (c == '/' && (flags & FNM_FILE_NAME))
		{
		  while (*n != '\0' && *n != '/')
		    ++n;
		  if (*n == '/'
		      && (internal_fnmatch (p, n + 1, flags & FNM_PERIOD,
					    flags) == 0))
		    return 0;
		}
	      else
		{
		  int flags2 = ((flags & FNM_FILE_NAME)
				? flags : (flags & ~FNM_PERIOD));

		  if (c == '\\' && !(flags & FNM_NOESCAPE))
		    c = *p;
		  c = FOLD (c);
		  for (--p; n < endp; ++n)
		    if (FOLD ((unsigned char) *n) == c
			&& (internal_fnmatch (p, n,
					      (no_leading_period
					       && (n == string
						   || (n[-1] == '/'
						       && (flags
							   & FNM_FILE_NAME)))),
					      flags2) == 0))
		      return 0;
		}
	    }

	  /* If we come here no match is possible with the wildcard.  */
	  return FNM_NOMATCH;

	case '[':
	  {
	    /* Nonzero if the sense of the character class is inverted.  */
	    static int posixly_correct;
	    register int not;
	    char cold;

	    if (posixly_correct == 0)
	      posixly_correct = getenv ("POSIXLY_CORRECT") != NULL ? 1 : -1;

	    if (*n == '\0')
	      return FNM_NOMATCH;

	    if (*n == '.' && no_leading_period && (n == string
						   || (n[-1] == '/'
						       && (flags
							   & FNM_FILE_NAME))))
	      return FNM_NOMATCH;

	    if (*n == '/' && (flags & FNM_FILE_NAME))
	      /* `/' cannot be matched.  */
	      return FNM_NOMATCH;

	    not = (*p == '!' || (posixly_correct < 0 && *p == '^'));
	    if (not)
	      ++p;

	    c = *p++;
	    for (;;)
	      {
		unsigned char fn = FOLD ((unsigned char) *n);

		if (!(flags & FNM_NOESCAPE) && c == '\\')
		  {
		    if (*p == '\0')
		      return FNM_NOMATCH;
		    c = FOLD ((unsigned char) *p);
		    ++p;

		    if (c == fn)
		      goto matched;
		  }
		else if (c == '[' && *p == ':')
		  {
		    /* Leave room for the null.  */
		    char str[CHAR_CLASS_MAX_LENGTH + 1];
		    size_t c1 = 0;
# if defined _LIBC || (defined HAVE_WCTYPE_H && defined HAVE_WCHAR_H)
		    wctype_t wt;
# endif
		    const char *startp = p;

		    for (;;)
		      {
			if (c1 > CHAR_CLASS_MAX_LENGTH)
			  /* The name is too long and therefore the pattern
			     is ill-formed.  */
			  return FNM_NOMATCH;

			c = *++p;
			if (c == ':' && p[1] == ']')
			  {
			    p += 2;
			    break;
			  }
			if (c < 'a' || c >= 'z')
			  {
			    /* This cannot possibly be a character class name.
			       Match it as a normal range.  */
			    p = startp;
			    c = '[';
			    goto normal_bracket;
			  }
			str[c1++] = c;
		      }
		    str[c1] = '\0';

# if defined _LIBC || (defined HAVE_WCTYPE_H && defined HAVE_WCHAR_H)
		    wt = IS_CHAR_CLASS (str);
		    if (wt == 0)
		      /* Invalid character class name.  */
		      return FNM_NOMATCH;

		    if (__iswctype (__btowc ((unsigned char) *n), wt))
		      goto matched;
# else
		    if ((STREQ (str, "alnum") && ISALNUM ((unsigned char) *n))
			|| (STREQ (str, "alpha") && ISALPHA ((unsigned char) *n))
			|| (STREQ (str, "blank") && ISBLANK ((unsigned char) *n))
			|| (STREQ (str, "cntrl") && ISCNTRL ((unsigned char) *n))
			|| (STREQ (str, "digit") && ISDIGIT ((unsigned char) *n))
			|| (STREQ (str, "graph") && ISGRAPH ((unsigned char) *n))
			|| (STREQ (str, "lower") && ISLOWER ((unsigned char) *n))
			|| (STREQ (str, "print") && ISPRINT ((unsigned char) *n))
			|| (STREQ (str, "punct") && ISPUNCT ((unsigned char) *n))
			|| (STREQ (str, "space") && ISSPACE ((unsigned char) *n))
			|| (STREQ (str, "upper") && ISUPPER ((unsigned char) *n))
			|| (STREQ (str, "xdigit") && ISXDIGIT ((unsigned char) *n)))
		      goto matched;
# endif
		  }
		else if (c == '\0')
		  /* [ (unterminated) loses.  */
		  return FNM_NOMATCH;
		else
		  {
		  normal_bracket:
		    if (FOLD (c) == fn)
		      goto matched;

		    cold = c;
		    c = *p++;

		    if (c == '-' && *p != ']')
		      {
			/* It is a range.  */
			unsigned char cend = *p++;
			if (!(flags & FNM_NOESCAPE) && cend == '\\')
			  cend = *p++;
			if (cend == '\0')
			  return FNM_NOMATCH;

			if (cold <= fn && fn <= FOLD (cend))
			  goto matched;

			c = *p++;
		      }
		  }

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

	    if (!not)
	      return FNM_NOMATCH;
	    break;

	  matched:
	    /* Skip the rest of the [...] that already matched.  */
	    while (c != ']')
	      {
		if (c == '\0')
		  /* [... (unterminated) loses.  */
		  return FNM_NOMATCH;

		c = *p++;
		if (!(flags & FNM_NOESCAPE) && c == '\\')
		  {
		    if (*p == '\0')
		      return FNM_NOMATCH;
		    /* XXX 1003.2d11 is unclear if this is right.  */
		    ++p;
		  }
		else if (c == '[' && *p == ':')
		  {
		    do
		      if (*++p == '\0')
			return FNM_NOMATCH;
		    while (*p != ':' || p[1] == ']');
		    p += 2;
		    c = *p;
		  }
	      }
	    if (not)
	      return FNM_NOMATCH;
	  }
	  break;

	default:
	  if (c != FOLD ((unsigned char) *n))
	    return FNM_NOMATCH;
	}

      ++n;
    }

  if (*n == '\0')
    return 0;

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

  return FNM_NOMATCH;

# undef FOLD
}