# -*- coding: utf-8 -*- # # BIFFwriter - An abstract base class for Excel workbooks and worksheets. # # # Used in conjunction with WriteExcel # # Copyright 2000-2010, John McNamara, jmcnamara@cpan.org # # original written in Perl by John McNamara # converted to Ruby by Hideo Nakamura, cxn03651@msj.biglobe.ne.jp # require 'tempfile' require 'writeexcel/write_file' class BIFFWriter < WriteFile #:nodoc: BIFF_Version = 0x0600 BigEndian = [1].pack("I") == [1].pack("N") attr_reader :byte_order, :data, :datasize ###################################################################### # The args here aren't used by BIFFWriter, but they are needed by its # subclasses. I don't feel like creating multiple constructors. ###################################################################### def initialize set_byte_order @data = '' @datasize = 0 @limit = 8224 @ignore_continue = 0 # Open a tmp file to store the majority of the Worksheet data. If this fails, # for example due to write permissions, store the data in memory. This can be # slow for large files. @filehandle = Tempfile.new('writeexcel') @filehandle.binmode # failed. store temporary data in memory. @using_tmpfile = @filehandle ? true : false end ############################################################################### # # _set_byte_order() # # Determine the byte order and store it as class data to avoid # recalculating it for each call to new(). # def set_byte_order # Check if "pack" gives the required IEEE 64bit float teststr = [1.2345].pack("d") hexdata = [0x8D, 0x97, 0x6E, 0x12, 0x83, 0xC0, 0xF3, 0x3F] number = hexdata.pack("C8") if number == teststr @byte_order = 0 # Little Endian elsif number == teststr.reverse @byte_order = 1 # Big Endian else # Give up. I'll fix this in a later version. raise( "Required floating point format not supported " + "on this platform. See the portability section " + "of the documentation." ) end end ############################################################################### # # get_data(). # # Retrieves data from memory in one chunk, or from disk in $buffer # sized chunks. # def get_data buflen = 4096 # Return data stored in memory unless @data.nil? tmp = @data @data = nil if @using_tmpfile @filehandle.open @filehandle.binmode end return tmp end # Return data stored on disk if @using_tmpfile return @filehandle.read(buflen) end # No data to return nil end ############################################################################### # # _store_bof($type) # # $type = 0x0005, Workbook # $type = 0x0010, Worksheet # $type = 0x0020, Chart # # Writes Excel BOF record to indicate the beginning of a stream or # sub-stream in the BIFF file. # def store_bof(type = 0x0005) record = 0x0809 # Record identifier length = 0x0010 # Number of bytes to follow # According to the SDK $build and $year should be set to zero. # However, this throws a warning in Excel 5. So, use these # magic numbers. build = 0x0DBB year = 0x07CC bfh = 0x00000041 sfo = 0x00000006 header = [record,length].pack("vv") data = [BIFF_Version,type,build,year,bfh,sfo].pack("vvvvVV") prepend(header, data) end ############################################################################### # # _store_eof() # # Writes Excel EOF record to indicate the end of a BIFF stream. # def store_eof record = 0x000A length = 0x0000 header = [record,length].pack("vv") append(header) end ############################################################################### # # _add_continue() # # Excel limits the size of BIFF records. In Excel 5 the limit is 2084 bytes. In # Excel 97 the limit is 8228 bytes. Records that are longer than these limits # must be split up into CONTINUE blocks. # # This function take a long BIFF record and inserts CONTINUE records as # necessary. # # Some records have their own specialised Continue blocks so there is also an # option to bypass this function. # def add_continue(data) record = 0x003C # Record identifier # Skip this if another method handles the continue blocks. return data if @ignore_continue != 0 # The first 2080/8224 bytes remain intact. However, we have to change # the length field of the record. # # in perl # $tmp = substr($data, 0, $limit, ""); if data.bytesize > @limit tmp = data[0, @limit] data[0, @limit] = '' else tmp = data.dup data = '' end tmp[2, 2] = [@limit-4].pack('v') # Strip out chunks of 2080/8224 bytes +4 for the header. while (data.bytesize > @limit) header = [record, @limit].pack("vv") tmp += header + data[0, @limit] data[0, @limit] = '' end # Mop up the last of the data header = [record, data.bytesize].pack("vv") tmp += header + data end ############################################################################### # # _add_mso_generic() # my $type = $_[0]; # my $version = $_[1]; # my $instance = $_[2]; # my $data = $_[3]; # # Create a mso structure that is part of an Escher drawing object. These are # are used for images, comments and filters. This generic method is used by # other methods to create specific mso records. # # Returns the packed record. # def add_mso_generic(type, version, instance, data, length = nil) length ||= data.bytesize # The header contains version and instance info packed into 2 bytes. header = version | (instance << 4) record = [header, type, length].pack('vvV') + data end def not_using_tmpfile @filehandle.close(true) if @filehandle @filehandle = nil @using_tmpfile = nil end def clear_data_for_test # :nodoc: @data = '' end def cleanup # :nodoc: @filehandle.close(true) if @filehandle end # override Object#inspect def inspect # :nodoc: to_s end end