httpparser.c

来自「原来由英特尔制定的UPnP SDK的」· C语言 代码 · 共 2,179 行 · 第 1/5 页

static UPNP_INLINE int
match_int( INOUT scanner_t * scanner,
           IN int base,
           OUT int *value )
{
    memptr token;
    token_type_t   tok_type;
    parse_status_t status;
    long           num;
    char          *end_ptr;
    size_t         save_pos;

    save_pos = scanner->cursor;

    status = scanner_get_token( scanner, &token, &tok_type );
    if( status == PARSE_OK ) {
        if( tok_type == TT_IDENTIFIER ) {
            errno = 0;

            num = strtol( token.buf, &end_ptr, base );
            if( ( num < 0 )
                // all and only those chars in token should be used for num
                || ( end_ptr != token.buf + token.length )
                || ( ( num == LONG_MIN || num == LONG_MAX )
                     && ( errno == ERANGE ) )
                 ) {
                status = PARSE_NO_MATCH;
            }

            *value = num;       // save result
        } else {
            status = PARSE_NO_MATCH;    // token must be an identifier
        }
    }

    if( status != PARSE_OK ) {
        // restore scanner position for bad values
        scanner->cursor = save_pos;
    }

    return status;
}

/************************************************************************
* Function: read_until_crlf
*
* Parameters:
*	INOUT scanner_t* scanner ;Scanner Object
*	OUT memptr* str ;	Buffer to copy scanner buffer contents to
*
* Description: Reads data until end of line; the crlf at the end of		
*	line is not consumed. On error, scanner is not restored. On			
*	success, 'str' points to a string that runs until eol				
*
* Returns:
*   PARSE_OK
*   PARSE_FAILURE
*   PARSE_INCOMPLETE
************************************************************************/
static UPNP_INLINE int
read_until_crlf( INOUT scanner_t * scanner,
                 OUT memptr * str )
{
    memptr token;
    token_type_t tok_type;
    parse_status_t status;
    size_t start_pos;

    start_pos = scanner->cursor;
    str->buf = scanner_get_str( scanner );

    // read until we hit a crlf
    do {
        status = scanner_get_token( scanner, &token, &tok_type );
    } while( status == PARSE_OK && tok_type != TT_CRLF );

    if( status == PARSE_OK ) {
        // pushback crlf in stream
        scanner->cursor -= token.length;

        // str should include all strings except crlf at the end
        str->length = scanner->cursor - start_pos;
    }

    return status;
}

/************************************************************************
* Function: skip_to_end_of_header
*
* Parameters:
*	INOUT scanner_t* scanner ; Scanner Object
*
* Description: Skip to end of header
*
* Returns:
*   PARSE_OK
*   PARSE_FAILURE
*   PARSE_INCOMPLETE
************************************************************************/
#ifndef WIN32
#warning There are currently no uses of the function 'skip_to_end_of_header()' in the code.
#warning 'skip_to_end_of_header()' is a candidate for removal.
#else
#pragma message("There are currently no uses of the function 'skip_to_end_of_header()' in the code.")
#pragma message("'skip_to_end_of_header()' is a candidate for removal.")
#endif
static UPNP_INLINE int
skip_to_end_of_header( INOUT scanner_t * scanner )
{
    memptr dummy_raw_value;
    parse_status_t status;

    status = match_raw_value( scanner, &dummy_raw_value );
    return status;
}

/************************************************************************
* Function: match_char
*
* Parameters:
*	INOUT scanner_t* scanner ;	Scanner Object
*	IN char c ;			Character to be compared with
*	IN xboolean case_sensitive;	Flag indicating whether
*					comparison should be case sensitive
*
* Description: Compares a character to the next char in the scanner;
*	on error, scanner chars are not restored
*
* Returns:
*   PARSE_OK
*   PARSE_NO_MATCH
*   PARSE_INCOMPLETE
************************************************************************/
static UPNP_INLINE parse_status_t
match_char( INOUT scanner_t * scanner,
            IN char c,
            IN xboolean case_sensitive )
{
    char scan_char;

    if( scanner->cursor >= scanner->msg->length ) {
        return PARSE_INCOMPLETE;
    }
    // read next char from scanner
    scan_char = scanner->msg->buf[scanner->cursor++];

    if( case_sensitive ) {
        return c == scan_char ? PARSE_OK : PARSE_NO_MATCH;
    } else {
        return tolower( c ) == tolower( scan_char ) ?
            PARSE_OK : PARSE_NO_MATCH;
    }
}

////////////////////////////////////////////////////////////////////////
// args for ...
//   %d,    int *     (31-bit positive integer)
//   %x,    int *     (31-bit postive number encoded as hex)
//   %s,    memptr*  (simple identifier)
//   %q,    memptr*  (quoted string)
//   %S,    memptr*  (non-whitespace string)
//   %R,    memptr*  (raw value)
//   %U,    uri_type* (url)
//   %L,    memptr*  (string until end of line)
//   %P,    int * (current index of the string being scanned)
//
// no args for
//   ' '    LWS*
//   \t     whitespace
//   "%%"   matches '%'
//   "% "   matches ' '
//   %c     matches CRLF
//   %i     ignore case in literal matching
//   %n     case-sensitive matching in literals
//   %w     optional whitespace; (similar to '\t', 
//                  except whitespace is optional)
//   %0     (zero) match null-terminator char '\0' 
//              (can only be used as last char in fmt)
//              use only in matchstr(), not match()
//   other chars match literally
//
// returns:
//   PARSE_OK
//   PARSE_INCOMPLETE
//   PARSE_FAILURE      -- bad input
//   PARSE_NO_MATCH     -- input does not match pattern

/************************************************************************
* Function : vfmatch
*
* Parameters :
* 	INOUT scanner_t* scanner ;	Scanner Object	
* 	IN const char* fmt ;		Pattern Format 
* 	va_list argp ;			List of variable arguments
*
* Description : Extracts variable parameters depending on the passed 
* 	in format parameter. Parses data also based on the passed in 
* 	format parameter.
*
* Return : int ;
* 	PARSE_OK
* 	PARSE_INCOMPLETE
* 	PARSE_FAILURE		- bad input
* 	PARSE_NO_MATCH		- input does not match pattern
*
* Note :
************************************************************************/
static int
vfmatch( INOUT scanner_t * scanner,
         IN const char *fmt,
         va_list argp )
{
    char c;
    const char *fmt_ptr = fmt;
    parse_status_t status;
    memptr *str_ptr;
    memptr temp_str;
    int *int_ptr;
    uri_type *uri_ptr;
    size_t save_pos;
    int stat;
    xboolean case_sensitive = TRUE;
    memptr token;
    token_type_t tok_type;
    int base;

    assert( scanner != NULL );
    assert( fmt != NULL );

    // save scanner pos; to aid error recovery
    save_pos = scanner->cursor;

    status = PARSE_OK;
    while( ( ( c = *fmt_ptr++ ) != 0 ) && ( status == PARSE_OK )
         ) {
        if( c == '%' ) {
            c = *fmt_ptr++;

            switch ( c ) {

                case 'R':      // raw value
                    str_ptr = va_arg( argp, memptr * );
                    assert( str_ptr != NULL );
                    status = match_raw_value( scanner, str_ptr );
                    break;

                case 's':      // simple identifier
                    str_ptr = va_arg( argp, memptr * );
                    assert( str_ptr != NULL );
                    status = scanner_get_token( scanner, str_ptr,
                                                &tok_type );
                    if( status == PARSE_OK && tok_type != TT_IDENTIFIER ) {
                        // not an identifier
                        status = PARSE_NO_MATCH;
                    }
                    break;

                case 'c':      // crlf
                    status = scanner_get_token( scanner,
                                                &token, &tok_type );
                    if( status == PARSE_OK && tok_type != TT_CRLF ) {
                        // not CRLF token
                        status = PARSE_NO_MATCH;
                    }
                    break;

                case 'd':      // integer
                case 'x':      // hex number
                    int_ptr = va_arg( argp, int * );

                    assert( int_ptr != NULL );
                    base = ( c == 'd' ? 10 : 16 );
                    status = match_int( scanner, base, int_ptr );
                    break;

                case 'S':      // non-whitespace string
                case 'U':      // uri
                    if( c == 'S' ) {
                        str_ptr = va_arg( argp, memptr * );
                    } else {
                        str_ptr = &temp_str;
                    }
                    assert( str_ptr != NULL );
                    status = match_non_ws_string( scanner, str_ptr );
                    if( c == 'U' && status == PARSE_OK ) {
                        uri_ptr = va_arg( argp, uri_type * );
                        assert( uri_ptr != NULL );
                        stat = parse_uri( str_ptr->buf, str_ptr->length,
                                          uri_ptr );
                        if( stat != HTTP_SUCCESS ) {
                            status = PARSE_NO_MATCH;
                        }
                    }
                    break;

                case 'L':      // string till eol
                    str_ptr = va_arg( argp, memptr * );
                    assert( str_ptr != NULL );
                    status = read_until_crlf( scanner, str_ptr );
                    break;

                case ' ':      // match space
                case '%':      // match percentage symbol
                    status = match_char( scanner, c, case_sensitive );
                    break;

                case 'n':      // case-sensitive match
                    case_sensitive = TRUE;
                    break;

                case 'i':      // ignore case
                    case_sensitive = FALSE;
                    break;

                case 'q':      // quoted string
                    str_ptr = ( memptr * ) va_arg( argp, memptr * );
                    status =
                        scanner_get_token( scanner, str_ptr, &tok_type );
                    if( status == PARSE_OK && tok_type != TT_QUOTEDSTRING ) {
                        status = PARSE_NO_MATCH;    // not a quoted string
                    }
                    break;

                case 'w':      // optional whitespace
                    status = scanner_get_token( scanner,
                                                &token, &tok_type );
                    if( status == PARSE_OK && tok_type != TT_WHITESPACE ) {
                        // restore non-whitespace token
                        scanner->cursor -= token.length;
                    }
                    break;

                case 'P':      // current pos of scanner
                    int_ptr = va_arg( argp, int * );

                    assert( int_ptr != NULL );
                    *int_ptr = scanner->cursor;
                    break;

                    // valid only in matchstr()
                case '0':      // end of msg?
                    // check that we are 1 beyond last char
                    if( scanner->cursor == scanner->msg->length &&
                        scanner->msg->buf[scanner->cursor] == '\0' ) {
                        status = PARSE_OK;
                    } else {
                        status = PARSE_NO_MATCH;
                    }
                    break;

                default:
                    assert( 0 );    // unknown option
            }
        } else {
            switch ( c ) {
                case ' ':      // LWS*
                    status = skip_lws( scanner );
                    break;

                case '\t':     // Whitespace
                    status = scanner_get_token( scanner,
                                                &token, &tok_type );
                    if( status == PARSE_OK && tok_type != TT_WHITESPACE ) {
                        // not whitespace token
                        status = PARSE_NO_MATCH;
                    }
                    break;

                default:       // match characters
                    {
                        status = match_char( scanner, c, case_sensitive );
                    }
            }
        }
    }

    if( status != PARSE_OK ) {
        // on error, restore original scanner pos
        scanner->cursor = save_pos;
    }

    return status;
}

/************************************************************************
* Function: match
*
* Parameters:
*	INOUT scanner_t* scanner ;	Scanner Object
*	IN const char* fmt;		Pattern format
*	...
*
* Description: matches a variable parameter list and takes necessary
*	actions based on the data type specified.
*
* Returns:
*   PARSE_OK
*   PARSE_NO_MATCH
*   PARSE_INCOMPLETE
************************************************************************/
static int
match( INOUT scanner_t * scanner,
       IN const char *fmt,
       ... )
{
    int ret_code;
    va_list args;

    va_start( args, fmt );
    ret_code = vfmatch( scanner, fmt, args );
    va_end( args );

    return ret_code;
}

/************************************************************************
* Function: matchstr
*
* Parameters:
*	IN char *str ;		String to be matched
*	IN size_t slen ;	Length of the string
*	IN const char* fmt ;	Pattern format
*	...
*
* Description: Matches a variable parameter list with a string
*	and takes actions based on the data type specified.
*
* Returns:
*   PARSE_OK
*   PARSE_NO_MATCH -- failure to match pattern 'fmt'
*   PARSE_FAILURE	-- 'str' is bad input
************************************************************************/
int
matchstr( IN char *str,
          IN size_t slen,