spreadsheet 1.3.3 → 1.3.4

data/lib/spreadsheet/excel/reader.rb

@@ -1,1314 +1,1360 @@
-require 'spreadsheet/encodings'
-require 'spreadsheet/font'
-require 'spreadsheet/formula'
-require 'spreadsheet/link'
-require 'spreadsheet/note'
-require 'spreadsheet/noteObject'
-require 'spreadsheet/excel/error'
-require 'spreadsheet/excel/internals'
-require 'spreadsheet/excel/sst_entry'
-require 'spreadsheet/excel/worksheet'
+require "spreadsheet/encodings"
+require "spreadsheet/font"
+require "spreadsheet/formula"
+require "spreadsheet/link"
+require "spreadsheet/note"
+require "spreadsheet/noteObject"
+require "spreadsheet/excel/error"
+require "spreadsheet/excel/internals"
+require "spreadsheet/excel/sst_entry"
+require "spreadsheet/excel/worksheet"
 
 module Spreadsheet
   module Excel
-##
-# Reader class for Excel Workbooks. Most read_* method correspond to an
-# Excel-Record/Opcode. You should not need to call any of its methods
-# directly. If you think you do, look at #read
-class Reader
-  include Spreadsheet::Encodings
-  include Spreadsheet::Excel::Internals
-  ROW_BLOCK_OPS = {
-    :blank  => true, :boolerr  => true, :dbcell   => true, :formula => true,
-    :label  => true, :labelsst => true, :mulblank => true, :mulrk   => true,
-    :number => true, :rk       => true, :rstring  => true,
-  }
-  def initialize opts = {}
-    @pos = 0
-    @bigendian = opts.fetch(:bigendian) {
-      [1].pack('l') != "\001\000\000\000"
-    }
-    @opts = opts
-    @boundsheets = nil
-    @current_row_block = {}
-    @current_row_block_offset = nil
-    @formats = {}
-    BUILTIN_FORMATS.each do |key, fmt| @formats.store key, client(fmt, 'UTF-8') end
-  end
-  def decode_rk work
-    #  Bit  Mask        Contents
-    #    0  0x00000001  0 = Value not changed 1 = Value is multiplied by 100
-    #    1  0x00000002  0 = Floating-point value 1 = Signed integer value
-    # 31-2  0xFFFFFFFC  Encoded value
-    #
-    # If bit 1 is cleared, the encoded value represents the 30 most significant
-    # bits of an IEEE 754 floating-point value (64-bit double precision). The
-    # 34 least significant bits must be set to zero. If bit 1 is set, the
-    # encoded value represents a signed 30-bit integer value. To get the
-    # correct integer, the encoded value has to be shifted right arithmetically
-    # by 2 bits. If bit 0 is set, the decoded value (both integer and
-    # floating-point) must be divided by 100 to get the final result.
-    flags, = work.unpack 'C'
-    cent = flags & 1
-    int  = flags & 2
-    value = 0
-    if int == 0
-      ## remove two bits
-      integer, = work.unpack 'V'
-      integer &= 0xfffffffc
-      value, = ("\0\0\0\0" +  [integer].pack('V')).unpack EIGHT_BYTE_DOUBLE
-    else
-      ## I can't find a format for unpacking a little endian signed integer.
-      #  'V' works for packing, but not for unpacking. But the following works
-      #  fine afaics:
-      unsigned, = (@bigendian ? work.reverse : work).unpack 'l'
-      ## remove two bits
-      value = unsigned >> 2
-    end
-    if cent == 1
-      value /= 100.0
-    end
-    value
-  end
-  def encoding codepage_id
-    name = CODEPAGES.fetch(codepage_id) do
-      raise Spreadsheet::Errors::UnknownCodepage, "Unknown Codepage 0x%04x" % codepage_id
-    end
+    ##
+    # Reader class for Excel Workbooks. Most read_* method correspond to an
+    # Excel-Record/Opcode. You should not need to call any of its methods
+    # directly. If you think you do, look at #read
+    class Reader
+      include Spreadsheet::Encodings
+      include Spreadsheet::Excel::Internals
+      ROW_BLOCK_OPS = {
+        blank: true, boolerr: true, dbcell: true, formula: true,
+        label: true, labelsst: true, mulblank: true, mulrk: true,
+        number: true, rk: true, rstring: true
+      }
+      def initialize opts = {}
+        @pos = 0
+        @bigendian = opts.fetch(:bigendian) {
+          [1].pack("l") != "\001\000\000\000"
+        }
+        @opts = opts
+        @boundsheets = nil
+        @current_row_block = {}
+        @current_row_block_offset = nil
+        @formats = {}
+        BUILTIN_FORMATS.each { |key, fmt| @formats.store key, client(fmt, "UTF-8") }
+      end
 
-    if RUBY_VERSION >= '1.9'
-      begin
-        Encoding.find name
-      rescue ArgumentError
-        raise Spreadsheet::Errors::UnsupportedEncoding, "Unsupported encoding with name '#{name}'"
+      def decode_rk work
+        #  Bit  Mask        Contents
+        #    0  0x00000001  0 = Value not changed 1 = Value is multiplied by 100
+        #    1  0x00000002  0 = Floating-point value 1 = Signed integer value
+        # 31-2  0xFFFFFFFC  Encoded value
+        #
+        # If bit 1 is cleared, the encoded value represents the 30 most significant
+        # bits of an IEEE 754 floating-point value (64-bit double precision). The
+        # 34 least significant bits must be set to zero. If bit 1 is set, the
+        # encoded value represents a signed 30-bit integer value. To get the
+        # correct integer, the encoded value has to be shifted right arithmetically
+        # by 2 bits. If bit 0 is set, the decoded value (both integer and
+        # floating-point) must be divided by 100 to get the final result.
+        flags, = work.unpack "C"
+        cent = flags & 1
+        int = flags & 2
+        value = 0
+        if int == 0
+          ## remove two bits
+          integer, = work.unpack "V"
+          integer &= 0xfffffffc
+          value, = ("\0\0\0\0" + [integer].pack("V")).unpack EIGHT_BYTE_DOUBLE
+        else
+          ## I can't find a format for unpacking a little endian signed integer.
+          #  'V' works for packing, but not for unpacking. But the following works
+          #  fine afaics:
+          unsigned, = (@bigendian ? work.reverse : work).unpack "l"
+          ## remove two bits
+          value = unsigned >> 2
+        end
+        if cent == 1
+          value /= 100.0
+        end
+        value
       end
-    else
-      name
-    end
-  end
-  def get_next_chunk
-    pos = @pos
-    if pos < @data.size
-      op, len = @data[@pos,OPCODE_SIZE].unpack('v2')
-      @pos += OPCODE_SIZE
-      if len
-        work = @data[@pos,len]
-        @pos += len
-        code = SEDOCPO.fetch(op, op)
-        if io = @opts[:print_opcodes]
-          io.puts sprintf("0x%04x/%-16s %5i: %s",
-                          op, code.inspect, len, work.inspect)
+
+      def encoding codepage_id
+        name = CODEPAGES.fetch(codepage_id) do
+          raise Spreadsheet::Errors::UnknownCodepage, "Unknown Codepage 0x%04x" % codepage_id
+        end
+
+        if RUBY_VERSION >= "1.9"
+          begin
+            Encoding.find name
+          rescue ArgumentError
+            raise Spreadsheet::Errors::UnsupportedEncoding, "Unsupported encoding with name '#{name}'"
+          end
+        else
+          name
         end
-        [ pos, code, len + OPCODE_SIZE, work]
       end
-    end
-  end
-  def in_row_block? op, previous
-    if op == :row
-      previous == op
-    else
-      ROW_BLOCK_OPS.include?(op)
-    end
-  end
-  def memoize?
-    @opts[:memoization]
-  end
-  def postread_workbook
-    sheets = @workbook.worksheets
-    sheets.each_with_index do |sheet, idx|
-      offset = sheet.offset
-      nxt = (nxtsheet = sheets[idx + 1]) ? nxtsheet.offset : @workbook.ole.size
-      @workbook.offsets.store sheet, [offset, nxt - offset]
-    end
-  end
-  def postread_worksheet worksheet
-     #We now have a lot of Note and NoteObjects, but they're not linked
-     #So link the noteObject(text) to the note (with author, position)
-     #TODO
-     @noteList.each do |i|
-        matching_objs = @noteObjList.select { |j| j.objID == i.objID }
-        if matching_objs.length > 1
-           puts "ERROR - more than one matching object ID!"
+
+      def get_next_chunk
+        pos = @pos
+        if pos < @data.size
+          op, len = @data[@pos, OPCODE_SIZE].unpack("v2")
+          @pos += OPCODE_SIZE
+          if len
+            work = @data[@pos, len]
+            @pos += len
+            code = SEDOCPO.fetch(op, op)
+            if (io = @opts[:print_opcodes])
+              io.puts sprintf("0x%04x/%-16s %5i: %s",
+                op, code.inspect, len, work.inspect)
+            end
+            [pos, code, len + OPCODE_SIZE, work]
+          end
         end
-        matching_obj = matching_objs.first
-        i.text = matching_obj.nil? ? '' : matching_obj.text
-        worksheet.add_note i.row, i.col, i.text
-     end
-  end
-  ##
-  # The entry-point for reading Excel-documents. Reads the Biff-Version and
-  # loads additional reader-methods before proceeding with parsing the document.
-  def read io
-    setup io
-    read_workbook
-    @workbook.default_format = @workbook.format 0
-    @workbook.changes.clear
-    @workbook
-  end
-  def read_blank worksheet, addr, work
-    # Offset  Size  Contents
-    #      0     2  Index to row
-    #      2     2  Index to column
-    #      4     2  Index to XF record (➜ 6.115)
-    row, column, xf = work.unpack binfmt(:blank)
-    set_cell worksheet, row, column, xf
-  end
-  def read_bof
-    # Offset  Size  Contents
-    #      0     2  BIFF version (always 0x0600 for BIFF8)
-    #      2     2  Type of the following data: 0x0005 = Workbook globals
-    #                                           0x0006 = Visual Basic module
-    #                                           0x0010 = Worksheet
-    #                                           0x0020 = Chart
-    #                                           0x0040 = Macro sheet
-    #                                           0x0100 = Workspace file
-    #      4     2  Build identifier
-    #      6     2  Build year
-    #      8     4  File history flags
-    #     12     4  Lowest Excel version that can read all records in this file
-    _, @bof, _, work = get_next_chunk
-    ## version and datatype are common to all Excel-Versions. Later versions
-    #  have additional information such as build-id and -year (from BIFF5).
-    #  These are ignored for the time being.
-    version, datatype = work.unpack('v2')
-    if datatype == 0x5
-      @version = version
-    end
-  end
-  def read_boolerr worksheet, addr, work
-    # Offset  Size  Contents
-    #      0     2  Index to row
-    #      2     2  Index to column
-    #      4     2  Index to XF record (➜ 6.115)
-    #      6     1  Boolean or error value (type depends on the following byte)
-    #      7     1  0 = Boolean value; 1 = Error code
-    row, column, xf, value, error = work.unpack 'v3C2'
-    set_cell worksheet, row, column, xf, error == 0 ? value > 0 : Error.new(value)
-  end
-  def read_boundsheet work, pos, len
-    # Offset  Size  Contents
-    #      0     4  Absolute stream position of the BOF record of the sheet
-    #               represented by this record. This field is never encrypted
-    #               in protected files.
-    #      4     1  Visibility: 0x00 = Visible
-    #                           0x01 = Hidden
-    #                           0x02 = Strong hidden (see below)
-    #      5     1  Sheet type: 0x00 = Worksheet
-    #                           0x02 = Chart
-    #                           0x06 = Visual Basic module
-    #      6  var.  Sheet name: BIFF5/BIFF7: Byte string,
-    #                           8-bit string length (➜ 3.3)
-    #                           BIFF8: Unicode string, 8-bit string length (➜ 3.4)
-    offset, visibility, _ = work.unpack("VC2")
-    name = client read_string(work[6..-1]), @workbook.encoding
-    if @boundsheets
-      @boundsheets[0] += 1
-      @boundsheets[2] += len
-    else
-      @boundsheets = [1, pos, len]
-    end
-    @workbook.set_boundsheets(*@boundsheets)
-    @workbook.add_worksheet Worksheet.new(:name     => name,
-                                          :ole      => @book,
-                                          :offset   => offset,
-                                          :reader   => self,
-                                          :visibility => WORKSHEET_VISIBILITIES[visibility])
-  end
-  def read_codepage work, pos, len
-    codepage, _ = work.unpack 'v'
-    @workbook.set_encoding encoding(codepage), pos, len
-  end
-  def read_colinfo worksheet, work, pos, len
-    # Offset  Size  Contents
-    #      0     2  Index to first column in the range
-    #      2     2  Index to last column in the range
-    #      4     2  Width of the columns in 1/256 of the width of the zero
-    #               character, using default font (first FONT record in the
-    #               file)
-    #      6     2  Index to XF record (➜ 6.115) for default column formatting
-    #      8     2  Option flags:
-    #               Bits  Mask    Contents
-    #                  0  0x0001  1 = Columns are hidden
-    #               10-8  0x0700  Outline level of the columns (0 = no outline)
-    #                 12  0x1000  1 = Columns are collapsed
-    #     10     2  Not used
-    first, last, width, xf, opts = work.unpack binfmt(:colinfo)[0..-2]
-    first.upto last do |col|
-      column = Column.new col, @workbook.format(xf),
-                          :width         => width.to_f / 256,
-                          :hidden        => (opts & 0x0001) > 0,
-                          :collapsed     => (opts & 0x1000) > 0,
-                          :outline_level => (opts & 0x0700) / 256
-      column.worksheet = worksheet
-      worksheet.columns[col] = column
-    end
-  end
-  def read_dimensions worksheet, work, pos, len
-    # Offset  Size  Contents
-    #      0     4  Index to first used row
-    #      4     4  Index to last used row, increased by 1
-    #      8     2  Index to first used column
-    #     10     2  Index to last used column, increased by 1
-    #     12     2  Not used
-    worksheet.set_dimensions work.unpack(binfmt(:dimensions)), pos, len
-  end
-  def read_font work, pos, len
-    # Offset  Size  Contents
-    #      0     2  Height of the font (in twips = 1/20 of a point)
-    #      2     2  Option flags:
-    #               Bit  Mask    Contents
-    #                 0  0x0001  1 = Characters are bold (redundant, see below)
-    #                 1  0x0002  1 = Characters are italic
-    #                 2  0x0004  1 = Characters are underlined
-    #                                (redundant, see below)
-    #                 3  0x0008  1 = Characters are struck out
-    #                 4  0x0010  1 = Characters are outlined (djberger)
-    #                 5  0x0020  1 = Characters are shadowed (djberger)
-    #      4     2  Colour index (➜ 6.70)
-    #      6     2  Font weight (100-1000). Standard values are
-    #                           0x0190 (400) for normal text and
-    #                           0x02bc (700) for bold text.
-    #      8     2  Escapement type: 0x0000 = None
-    #                                0x0001 = Superscript
-    #                                0x0002 = Subscript
-    #     10     1  Underline type: 0x00 = None
-    #                               0x01 = Single
-    #                               0x02 = Double
-    #                               0x21 = Single accounting
-    #                               0x22 = Double accounting
-    #     11     1  Font family:
-    #               0x00 = None (unknown or don't care)
-    #               0x01 = Roman (variable width, serifed)
-    #               0x02 = Swiss (variable width, sans-serifed)
-    #               0x03 = Modern (fixed width, serifed or sans-serifed)
-    #               0x04 = Script (cursive)
-    #               0x05 = Decorative (specialised,
-    #                                  for example Old English, Fraktur)
-    #     12     1  Character set: 0x00 =   0 = ANSI Latin
-    #                              0x01 =   1 = System default
-    #                              0x02 =   2 = Symbol
-    #                              0x4d =  77 = Apple Roman
-    #                              0x80 = 128 = ANSI Japanese Shift-JIS
-    #                              0x81 = 129 = ANSI Korean (Hangul)
-    #                              0x82 = 130 = ANSI Korean (Johab)
-    #                              0x86 = 134 = ANSI Chinese Simplified GBK
-    #                              0x88 = 136 = ANSI Chinese Traditional BIG5
-    #                              0xa1 = 161 = ANSI Greek
-    #                              0xa2 = 162 = ANSI Turkish
-    #                              0xa3 = 163 = ANSI Vietnamese
-    #                              0xb1 = 177 = ANSI Hebrew
-    #                              0xb2 = 178 = ANSI Arabic
-    #                              0xba = 186 = ANSI Baltic
-    #                              0xcc = 204 = ANSI Cyrillic
-    #                              0xde = 222 = ANSI Thai
-    #                              0xee = 238 = ANSI Latin II (Central European)
-    #                              0xff = 255 = OEM Latin I
-    #     13     1  Not used
-    #     14  var.  Font name:
-    #               BIFF5/BIFF7: Byte string, 8-bit string length (➜ 3.3)
-    #               BIFF8: Unicode string, 8-bit string length (➜ 3.4)
-    name = client read_string(work[14..-1]), @workbook.encoding
-    font = Font.new name
-    size, opts, color, font.weight, escapement, underline,
-      family, encoding = work.unpack binfmt(:font)
-    font.size       = size / TWIPS
-    font.italic     = opts & 0x0002
-    font.strikeout  = opts & 0x0008
-    font.outline    = opts & 0x0010
-    font.shadow     = opts & 0x0020
-    font.color      = COLOR_CODES[color] || :text
-    font.escapement = ESCAPEMENT_TYPES[escapement]
-    font.underline  = UNDERLINE_TYPES[underline]
-    font.family     = FONT_FAMILIES[family]
-    font.encoding   = FONT_ENCODINGS[encoding]
-    @workbook.add_font font
-  end
-  def read_format work, pos, len
-    # Offset  Size  Contents
-    #      0     2  Format index used in other records
-    #      2  var.  Number format string
-    #               (Unicode string, 16-bit string length, ➜ 3.4)
-    idx, = work.unpack 'v'
-    value = read_string work[2..-1], 2
-    @formats.store idx, client(value, @workbook.encoding)
-  end
-  def read_formula worksheet, addr, work
-    # Offset  Size  Contents
-    #      0     2  Index to row
-    #      2     2  Index to column
-    #      4     2  Index to XF record (➜ 6.115)
-    #      6     8  Result of the formula. See below for details.
-    #     14     2  Option flags:
-    #               Bit  Mask    Contents
-    #                 0  0x0001  1 = Recalculate always
-    #                 1  0x0002  1 = Calculate on open
-    #                 3  0x0008  1 = Part of a shared formula
-    #     16     4  Not used
-    #     20  var.  Formula data (RPN token array, ➜ 4)
-    #               Offset  Size  Contents
-    #                    0     2  Size of the following formula data (sz)
-    #                    2    sz  Formula data (RPN token array)
-    #               [2+sz]  var.  (optional) Additional data for specific tokens
-    #                             (➜ 4.1.6, for example tArray token, ➜ 4.8.7)
-    #
-    # Result of the Formula
-    # Dependent on the type of value the formula returns, the result field has
-    # the following format:
-    #
-    # Result is a numeric value:
-    # Offset  Size  Contents
-    #      0     8  IEEE 754 floating-point value (64-bit double precision)
-    #
-    # Result is a string (the string follows in a STRING record, ➜ 6.98):
-    # Offset  Size  Contents
-    #      0     1  0x00 (identifier for a string value)
-    #      1     5  Not used
-    #      6     2  0xffff
-    # Note: In BIFF8 the string must not be empty. For empty cells there is a
-    # special identifier defined (see below).
-    #
-    # Result is a Boolean value:
-    # Offset  Size  Contents
-    #      0     1  0x01 (identifier for a Boolean value)
-    #      1     1  Not used
-    #      2     1  0 = FALSE, 1 = TRUE
-    #      3     3  Not used
-    #      6     2  0xffff
-    #
-    # Result is an error value:
-    # Offset  Size  Contents
-    #      0     1  0x02 (identifier for an error value)
-    #      1     1  Not used
-    #      2     1  Error code (➜ 3.7)
-    #      3     3  Not used
-    #      6     2  0xffff
-    #
-    # Result is an empty cell (BIFF8), for example an empty string:
-    # Offset  Size  Contents
-    #      0     1  0x03 (identifier for an empty cell)
-    #      1     5  Not used
-    #      6     2  0xffff
-    row, column, xf, rtype, rval, rcheck, opts = work.unpack 'v3CxCx3v2'
-    formula = Formula.new
-    formula.shared = (opts & 0x08) > 0
-    formula.data = work[20..-1]
-    if rcheck != 0xffff || rtype > 3
-      value, = work.unpack 'x6E'
-      unless value
-        # on architectures where sizeof(double) > 8
-        value, = work.unpack 'x6e'
       end
-      formula.value = value
-    elsif rtype == 0
-      pos, op, _len, work = get_next_chunk
-      if op == :sharedfmla
-        ## TODO: formula-support in 0.8.0
-        pos, op, _len, work = get_next_chunk
+
+      def in_row_block? op, previous
+        if op == :row
+          previous == op
+        else
+          ROW_BLOCK_OPS.include?(op)
+        end
       end
-      if op == :string
-        formula.value = client read_string(work, 2), @workbook.encoding
-      else
-        warn "String Value expected after Formula, but got #{op}"
-        formula.value = Error.new 0x2a
-        @pos = pos
+
+      def memoize?
+        @opts[:memoization]
       end
-    elsif rtype == 1
-      formula.value = rval > 0
-    elsif rtype == 2
-      formula.value = Error.new rval
-    else
-      # leave the Formula value blank
-    end
-    set_cell worksheet, row, column, xf, formula
-  end
-  def read_hlink worksheet, work, pos, len
-    # 6.53.1 Common Record Contents
-    # Offset  Size  Contents
-    #      0     8  Cell range address of all cells containing this hyperlink
-    #               (➜ 3.13.1)
-    #      8    16  GUID of StdLink:
-    #               D0 C9 EA 79 F9 BA CE 11 8C 82 00 AA 00 4B A9 0B
-    #               (79EAC9D0-BAF9-11CE-8C82-00AA004BA90B)
-    #     24     4  Unknown value: 0x00000002
-    #     28     4  Option flags (see below)
-    #               Bit  Mask        Contents
-    #                 0  0x00000001  0 = No link extant
-    #                                1 = File link or URL
-    #                 1  0x00000002  0 = Relative file path
-    #                                1 = Absolute path or URL
-    #           2 and 4  0x00000014  0 = No description
-    #                                1 (both bits) = Description
-    #                 3  0x00000008  0 = No text mark
-    #                                1 = Text mark
-    #                 7  0x00000080  0 = No target frame
-    #                                1 = Target frame
-    #                 8  0x00000100  0 = File link or URL
-    #                                1 = UNC path (incl. server name)
-    #--------------------------------------------------------------------------
-    #   [32]     4  (optional, see option flags) Character count of description
-    #               text, including trailing zero word (dl)
-    #   [36]  2∙dl  (optional, see option flags) Character array of description
-    #               text, no Unicode string header, always 16-bit characters,
-    #               zero-terminated
-    #--------------------------------------------------------------------------
-    # [var.]     4  (optional, see option flags) Character count of target
-    #               frame, including trailing zero word (fl)
-    # [var.]  2∙fl  (optional, see option flags) Character array of target
-    #               frame, no Unicode string header, always 16-bit characters,
-    #               zero-terminated
-    #--------------------------------------------------------------------------
-    #   var.  var.  Special data (➜ 6.53.2 and following)
-    #--------------------------------------------------------------------------
-    # [var.]     4  (optional, see option flags) Character count of the text
-    #               mark, including trailing zero word (tl)
-    # [var.]  2∙tl  (optional, see option flags) Character array of the text
-    #               mark without “#” sign, no Unicode string header, always
-    #               16-bit characters, zero-terminated
-    firstrow, lastrow, firstcol, lastcol, _, opts = work.unpack 'v4H32x4V'
-    has_link = opts & 0x0001
-    desc     = opts & 0x0014
-    textmark = opts & 0x0008
-    target   = opts & 0x0080
-    unc      = opts & 0x0100
-    link = Link.new
-    _, description = nil
-    pos = 32
-    if desc > 0
-      description, pos = read_hlink_string work, pos
-      link << description
-    end
-    if target > 0
-      link.target_frame, pos = read_hlink_string work, pos
-    end
-    if unc > 0
-      # 6.53.4 Hyperlink to a File with UNC (Universal Naming Convention) Path
-      # These data fields are for UNC paths containing a server name (for
-      # instance “\\server\path\file.xls”). The lower 9 bits of the option
-      # flags field must be 1.x00x.xx112.
-      # Offset  Size  Contents
-      #      0     4  Character count of the UNC,
-      #               including trailing zero word (fl)
-      #      4  2∙fl  Character array of the UNC, no Unicode string header,
-      #               always 16-bit characters, zeroterminated.
-      link.url, pos = read_hlink_string work, pos
-    elsif has_link > 0
-      uid, = work.unpack "x#{pos}H32"
-      pos += 16
-      if uid == "e0c9ea79f9bace118c8200aa004ba90b"
-        # 6.53.2 Hyperlink containing a URL (Uniform Resource Locator)
-        # These data fields occur for links which are not local files or files
-        # in the local network (for instance HTTP and FTP links and e-mail
-        # addresses). The lower 9 bits of the option flags field must be
-        # 0.x00x.xx112 (x means optional, depending on hyperlink content). The
-        # GUID could be used to distinguish a URL from a file link.
+
+      def postread_workbook
+        sheets = @workbook.worksheets
+        sheets.each_with_index do |sheet, idx|
+          offset = sheet.offset
+          nxt = (nxtsheet = sheets[idx + 1]) ? nxtsheet.offset : @workbook.ole.size
+          @workbook.offsets.store sheet, [offset, nxt - offset]
+        end
+      end
+
+      def postread_worksheet worksheet
+        # We now have a lot of Note and NoteObjects, but they're not linked
+        # So link the noteObject(text) to the note (with author, position)
+        # TODO
+        @note_list.each do |i|
+          matching_objs = @note_ob_list.select { |j| j.obj_id == i.obj_id }
+          if matching_objs.length > 1
+            puts "ERROR - more than one matching object ID!"
+          end
+          matching_obj = matching_objs.first
+          i.text = matching_obj.nil? ? "" : matching_obj.text
+          worksheet.add_note i.row, i.col, i.text
+        end
+      end
+
+      ##
+      # The entry-point for reading Excel-documents. Reads the Biff-Version and
+      # loads additional reader-methods before proceeding with parsing the document.
+      def read io
+        setup io
+        read_workbook
+        @workbook.default_format = @workbook.format 0
+        @workbook.changes.clear
+        @workbook
+      end
+
+      def read_blank worksheet, addr, work
         # Offset  Size  Contents
-        #      0    16  GUID of URL Moniker:
-        #               E0 C9 EA 79 F9 BA CE 11 8C 82 00 AA 00 4B A9 0B
-        #               (79EAC9E0-BAF9-11CE-8C82-00AA004BA90B)
-        #     16     4  Size of character array of the URL, including trailing
-        #               zero word (us). There are us/2-1 characters in the
-        #               following string.
-        #     20    us  Character array of the URL, no Unicode string header,
-        #               always 16-bit characters, zeroterminated
-        size, = work.unpack "x#{pos}V"
-        pos += 4
-        data = work[pos, size].chomp "\000\000"
-        link.url = client data
-        pos += size
-      else
-        # 6.53.3 Hyperlink to a Local File
-        # These data fields are for links to files on local drives. The path of
-        # the file can be complete with drive letter (absolute) or relative to
-        # the location of the workbook. The lower 9 bits of the option flags
-        # field must be 0.x00x.xxx12. The GUID could be used to distinguish a
-        # URL from a file link.
+        #      0     2  Index to row
+        #      2     2  Index to column
+        #      4     2  Index to XF record (➜ 6.115)
+        row, column, xf = work.unpack binfmt(:blank)
+        set_cell worksheet, row, column, xf
+      end
+
+      def read_bof
         # Offset  Size  Contents
-        #      0    16  GUID of File Moniker:
-        #               03 03 00 00 00 00 00 00 C0 00 00 00 00 00 00 46
-        #               (00000303-0000-0000-C000-000000000046)
-        #     16     2  Directory up-level count. Each leading “..\” in the
-        #               file link is deleted and increases this counter.
-        #     18     4  Character count of the shortened file path and name,
-        #               including trailing zero byte (sl)
-        #     22    sl  Character array of the shortened file path and name in
-        #               8.3-DOS-format. This field can be filled with a long
-        #               file name too. No Unicode string header, always 8-bit
-        #               characters, zeroterminated.
-        #  22+sl    24  Unknown byte sequence:
-        #               FF FF AD DE 00 00 00 00
-        #               00 00 00 00 00 00 00 00
-        #               00 00 00 00 00 00 00 00
-        #  46+sl     4  Size of the following file link field including string
-        #               length field and additional data field (sz). If sz is
-        #               zero, nothing will follow (except a text mark).
-        # [50+sl]    4  (optional) Size of character array of the extended file
-        #               path and name (xl). There are xl/2 characters in the
-        #               following string.
-        # [54+sl]    2  (optional) Unknown byte sequence: 03 00
-        # [56+sl]   xl  (optional) Character array of the extended file path
-        #               and name (xl), no Unicode string header, always 16-bit
-        #               characters, not zero-terminated
-        uplevel, count = work.unpack "x#{pos}vV"
-        pos += 6
-        # TODO: short file path may have any of the OEM encodings. Find out which
-        #       and use the #client method to convert the encoding.
-        prefix = internal('..\\', 'UTF-8') * uplevel
-        link.dos = link.url = prefix << work[pos, count].chomp("\000")
-        pos += count + 24
-        total, size = work.unpack "x#{pos}V2"
-        pos += 10
-        if total > 0
-          link.url = client work[pos, size]
-          pos += size
+        #      0     2  BIFF version (always 0x0600 for BIFF8)
+        #      2     2  Type of the following data: 0x0005 = Workbook globals
+        #                                           0x0006 = Visual Basic module
+        #                                           0x0010 = Worksheet
+        #                                           0x0020 = Chart
+        #                                           0x0040 = Macro sheet
+        #                                           0x0100 = Workspace file
+        #      4     2  Build identifier
+        #      6     2  Build year
+        #      8     4  File history flags
+        #     12     4  Lowest Excel version that can read all records in this file
+        _, @bof, _, work = get_next_chunk
+        ## version and datatype are common to all Excel-Versions. Later versions
+        #  have additional information such as build-id and -year (from BIFF5).
+        #  These are ignored for the time being.
+        version, datatype = work.unpack("v2")
+        if datatype == 0x5
+          @version = version
         end
       end
-    else
-      # 6.53.5 Hyperlink to the Current Workbook
-      # In this case only the text mark field is present (optional with
-      # description).
-      # Example: The URL “#Sheet2!B1:C2” refers to the given range in the
-      # current workbook.
-      # The lower 9 bits of the option flags field must be 0.x00x.1x002.
-    end
-    if textmark > 0
-      link.fragment, _ = read_hlink_string work, pos
-    end
-    if link.empty?
-      link << link.href
-    end
-    firstrow.upto lastrow do |row|
-      firstcol.upto lastcol do |col|
-        worksheet.add_link row, col, link
+
+      def read_boolerr worksheet, addr, work
+        # Offset  Size  Contents
+        #      0     2  Index to row
+        #      2     2  Index to column
+        #      4     2  Index to XF record (➜ 6.115)
+        #      6     1  Boolean or error value (type depends on the following byte)
+        #      7     1  0 = Boolean value; 1 = Error code
+        row, column, xf, value, error = work.unpack "v3C2"
+        set_cell worksheet, row, column, xf, (error == 0) ? value > 0 : Error.new(value)
       end
-    end
-  end
-  def read_hlink_string work, pos
-    count, = work.unpack "x#{pos}V"
-    len = count * 2
-    pos += 4
-    data = work[pos, len].chomp "\000\000"
-    pos += len
-    [client(data, 'UTF-16LE'), pos]
-  end
-  def read_index worksheet, work, pos, len
-    # Offset  Size  Contents
-    #      0     4  Not used
-    #      4     4  Index to first used row (rf, 0-based)
-    #      8     4  Index to first row of unused tail of sheet
-    #               (rl, last used row + 1, 0-based)
-    #     12     4  Absolute stream position of the
-    #               DEFCOLWIDTH record (➜ 6.29) of the current sheet. If this
-    #               record does not exist, the offset points to the record at
-    #               the position where the DEFCOLWIDTH record would occur.
-    #     16  4∙nm  Array of nm absolute stream positions to the
-    #               DBCELL record (➜ 6.26) of each Row Block
-    # TODO: use the index if it exists
-    # _, first_used, first_unused, defcolwidth, *indices = work.unpack 'V*'
-  end
-  def read_label worksheet, addr, work
-    # Offset  Size  Contents
-    #      0     2  Index to row
-    #      2     2  Index to column
-    #      4     2  Index to XF record (➜ 6.115)
-    #      6  var.  Unicode string, 16-bit string length (➜ 3.4)
-    row, column, xf = work.unpack 'v3'
-    value = client read_string(work[6..-1], 2), @workbook.encoding
-    set_cell worksheet, row, column, xf, value
-  end
-  def read_labelsst worksheet, addr, work
-    # Offset  Size  Contents
-    #      0     2  Index to row
-    #      2     2  Index to column
-    #      4     2  Index to XF record (➜ 6.115)
-    #      6     4  Index into SST record (➜ 6.96)
-    row, column, xf, index = work.unpack binfmt(:labelsst)
-    set_cell worksheet, row, column, xf, worksheet.shared_string(index)
-  end
-  def read_mulblank worksheet, addr, work
-    # Offset  Size  Contents
-    #      0     2  Index to row
-    #      2     2  Index to first column (fc)
-    #      4  2∙nc  List of nc=lc-fc+1 16-bit indexes to XF records (➜ 6.115)
-    # 4+2∙nc     2  Index to last column (lc)
-    row, column, *xfs = work.unpack 'v*'
-    xfs.pop #=> last_column
-    xfs.each_with_index do |xf, idx| set_cell worksheet, row, column + idx, xf end
-  end
-  def read_mulrk worksheet, addr, work
-    # Offset  Size  Contents
-    #      0     2  Index to row
-    #      2     2  Index to first column (fc)
-    #      4  6∙nc  List of nc=lc-fc+1 XF/RK structures. Each XF/RK contains:
-    #               Offset  Size  Contents
-    #                    0     2  Index to XF record (➜ 6.115)
-    #                    2     4  RK value (➜ 3.6)
-    # 4+6∙nc     2  Index to last column (lc)
-    row, column = work.unpack 'v2'
-    4.step(work.size - 6, 6) do |idx|
-      xf, = work.unpack "x#{idx}v"
-      set_cell worksheet, row, column, xf, decode_rk(work[idx + 2, 4])
-      column += 1
-    end
-  end
-  def read_number worksheet, addr, work
-    # Offset  Size  Contents
-    #      0     2  Index to row
-    #      2     2  Index to column
-    #      4     2  Index to XF record (➜ 6.115)
-    #      6     8  IEEE 754 floating-point value (64-bit double precision)
-    row, column, xf, value = work.unpack binfmt(:number)
-    set_cell worksheet, row, column, xf, value
-  end
-  def read_rk worksheet, addr, work
-    # Offset  Size  Contents
-    #      0     2  Index to row
-    #      2     2  Index to column
-    #      4     2  Index to XF record (➜ 6.115)
-    #      6     4  RK value (➜ 3.6)
-    row, column, xf = work.unpack 'v3'
-    set_cell worksheet, row, column, xf, decode_rk(work[6,4])
-  end
-  def read_row worksheet, addr
-    row = addr[:index]
-    @current_row_block.fetch [worksheet, row] do
-      @current_row_block.clear
-      cells = @current_row_block[[worksheet, row]] = Row.new(nil, row)
-      @pos = addr[:offset]
-      found = false
-      while tuple = get_next_chunk
-        pos, op, _, work = tuple
-        case op
-        when :eof      # ●  EOF ➜ 6.36 - we should only get here if there is just
-                       #                 one Row-Block
-          @pos = pos
-          return cells
-        when :dbcell   # ○  DBCELL Stream offsets to the cell records of each row
-          return cells
-        when :row      # ○○ Row Blocks ➜ 5.7
-                       # ●  ROW ➜ 6.83
-          # ignore, we already did these in read_worksheet
-          return cells if found
-        when :blank    # BLANK ➜ 6.7
-          found = true
-          read_blank worksheet, addr, work
-        when :boolerr  # BOOLERR ➜ 6.10
-          found = true
-          read_boolerr worksheet, addr, work
-        when 0x0002    # INTEGER ➜ 6.56 (BIFF2 only)
-          found = true
-          # TODO: implement for BIFF2 support
-        when :formula  # FORMULA ➜ 6.46
-          found = true
-          read_formula worksheet, addr, work
-        when :label    # LABEL ➜ 6.59 (BIFF2-BIFF7)
-          found = true
-          read_label worksheet, addr, work
-        when :labelsst # LABELSST ➜ 6.61 (BIFF8 only)
-          found = true
-          read_labelsst worksheet, addr, work
-        when :mulblank # MULBLANK ➜ 6.64 (BIFF5-BIFF8)
-          found = true
-          read_mulblank worksheet, addr, work
-        when :mulrk    # MULRK ➜ 6.65 (BIFF5-BIFF8)
-          found = true
-          read_mulrk worksheet, addr, work
-        when :number   # NUMBER ➜ 6.68
-          found = true
-          read_number worksheet, addr, work
-        when :rk       # RK ➜ 6.82 (BIFF3-BIFF8)
-          found = true
-          read_rk worksheet, addr, work
-        when :rstring  # RSTRING ➜ 6.84 (BIFF5/BIFF7)
-          found = true
-          read_rstring worksheet, addr, work
+
+      def read_boundsheet work, pos, len
+        # Offset  Size  Contents
+        #      0     4  Absolute stream position of the BOF record of the sheet
+        #               represented by this record. This field is never encrypted
+        #               in protected files.
+        #      4     1  Visibility: 0x00 = Visible
+        #                           0x01 = Hidden
+        #                           0x02 = Strong hidden (see below)
+        #      5     1  Sheet type: 0x00 = Worksheet
+        #                           0x02 = Chart
+        #                           0x06 = Visual Basic module
+        #      6  var.  Sheet name: BIFF5/BIFF7: Byte string,
+        #                           8-bit string length (➜ 3.3)
+        #                           BIFF8: Unicode string, 8-bit string length (➜ 3.4)
+        offset, visibility, _ = work.unpack("VC2")
+        name = client read_string(work[6..]), @workbook.encoding
+        if @boundsheets
+          @boundsheets[0] += 1
+          @boundsheets[2] += len
+        else
+          @boundsheets = [1, pos, len]
         end
+        @workbook.set_boundsheets(*@boundsheets)
+        @workbook.add_worksheet Worksheet.new(name: name,
+          ole: @book,
+          offset: offset,
+          reader: self,
+          visibility: WORKSHEET_VISIBILITIES[visibility])
       end
-      cells
-    end
-  end
-  def read_rstring worksheet, addr, work
-    # Offset  Size  Contents
-    #      0     2  Index to row
-    #      2     2  Index to column
-    #      4     2  Index to XF record (➜ 6.115)
-    #      6    sz  Unformatted Unicode string, 16-bit string length (➜ 3.4)
-    #   6+sz     2  Number of Rich-Text formatting runs (rt)
-    #   8+sz  4·rt  List of rt formatting runs (➜ 3.2)
-    row, column, xf = work.unpack 'v3'
-    value = client read_string(work[6..-1], 2), @workbook.encoding
-    set_cell worksheet, row, column, xf, value
-  end
-  def read_window2 worksheet, work, pos, len
-    # This record contains additional settings for the document window
-    # (BIFF2-BIFF4) or for the window of a specific worksheet (BIFF5-BIFF8).
-    # It is part of the Sheet View Settings Block (➜ 4.5).
-    # Offset  Size  Contents
-    #      0     2  Option flags:
-    #               Bits  Mask    Contents
-    #                  0  0x0001  0 = Show formula results
-    #                             1 = Show formulas
-    #                  1  0x0002  0 = Do not show grid lines
-    #                             1 = Show grid lines
-    #                  2  0x0004  0 = Do not show sheet headers
-    #                             1 = Show sheet headers
-    #                  3  0x0008  0 = Panes are not frozen
-    #                             1 = Panes are frozen (freeze)
-    #                  4  0x0010  0 = Show zero values as empty cells
-    #                             1 = Show zero values
-    #                  5  0x0020  0 = Manual grid line colour
-    #                             1 = Automatic grid line colour
-    #                  6  0x0040  0 = Columns from left to right
-    #                             1 = Columns from right to left
-    #                  7  0x0080  0 = Do not show outline symbols
-    #                             1 = Show outline symbols
-    #                  8  0x0100  0 = Keep splits if pane freeze is removed
-    #                             1 = Remove splits if pane freeze is removed
-    #                  9  0x0200  0 = Sheet not selected
-    #                             1 = Sheet selected (BIFF5-BIFF8)
-    #                 10  0x0400  0 = Sheet not active
-    #                             1 = Sheet active (BIFF5-BIFF8)
-    #                 11  0x0800  0 = Show in normal view
-    #                             1 = Show in page break preview (BIFF8)
-    #      2     2  Index to first visible row
-    #      4     2  Index to first visible column
-    #      6     2  Colour index of grid line colour (➜ 5.74).
-    #               Note that in BIFF2-BIFF5 an RGB colour is written instead.
-    #      8     2  Not used
-    #     10     2  Cached magnification factor in page break preview (in percent)
-    #               0 = Default (60%)
-    #     12     2  Cached magnification factor in normal view (in percent)
-    #               0 = Default (100%)
-    #     14     4  Not used
-    flags, _ = work.unpack 'v'
-    worksheet.selected = flags & 0x0200 > 0
-  end
 
-  def read_merged_cells worksheet, work, pos, len
-    # This record contains the addresses of merged cell ranges in the current sheet.
-    # Record MERGEDCELLS, BIFF8:
-    # Offset  Size  Contents
-    # 0	      var.	Cell range address list with merged ranges (➜ 2.5.15)
-    # If the record size exceeds the limit, it is not continued with a CONTINUE record,
-    # but another self-contained MERGEDCELLS record is started. The limit of 8224 bytes
-    # per record results in a maximum number of 1027 merged ranges.
+      def read_codepage work, pos, len
+        codepage, _ = work.unpack "v"
+        @workbook.set_encoding encoding(codepage), pos, len
+      end
 
-    worksheet.merged_cells.push(*read_range_address_list(work, len))
-    #
-    # A cell range address list consists of a field with the number of ranges and the list
-    # of the range addresses.
-    # Cell range address list, BIFF2-BIFF8:
-    # Offset  Size          Contents
-    # 0       2             Number of following cell range addresses (nm)
-    # 2	      6∙nm or 8∙nm  List of nm cell range addresses (➜ 2.5.14)
-    #
-  end
+      def read_colinfo worksheet, work, pos, len
+        # Offset  Size  Contents
+        #      0     2  Index to first column in the range
+        #      2     2  Index to last column in the range
+        #      4     2  Width of the columns in 1/256 of the width of the zero
+        #               character, using default font (first FONT record in the
+        #               file)
+        #      6     2  Index to XF record (➜ 6.115) for default column formatting
+        #      8     2  Option flags:
+        #               Bits  Mask    Contents
+        #                  0  0x0001  1 = Columns are hidden
+        #               10-8  0x0700  Outline level of the columns (0 = no outline)
+        #                 12  0x1000  1 = Columns are collapsed
+        #     10     2  Not used
+        first, last, width, xf, opts = work.unpack binfmt(:colinfo)[0..-2]
+        first.upto last do |col|
+          column = Column.new col, @workbook.format(xf),
+            width: width.to_f / 256,
+            hidden: (opts & 0x0001) > 0,
+            collapsed: (opts & 0x1000) > 0,
+            outline_level: (opts & 0x0700) / 256
+          column.worksheet = worksheet
+          worksheet.columns[col] = column
+        end
+      end
+
+      def read_dimensions worksheet, work, pos, len
+        # Offset  Size  Contents
+        #      0     4  Index to first used row
+        #      4     4  Index to last used row, increased by 1
+        #      8     2  Index to first used column
+        #     10     2  Index to last used column, increased by 1
+        #     12     2  Not used
+        worksheet.set_dimensions work.unpack(binfmt(:dimensions)), pos, len
+      end
+
+      def read_font work, pos, len
+        # Offset  Size  Contents
+        #      0     2  Height of the font (in twips = 1/20 of a point)
+        #      2     2  Option flags:
+        #               Bit  Mask    Contents
+        #                 0  0x0001  1 = Characters are bold (redundant, see below)
+        #                 1  0x0002  1 = Characters are italic
+        #                 2  0x0004  1 = Characters are underlined
+        #                                (redundant, see below)
+        #                 3  0x0008  1 = Characters are struck out
+        #                 4  0x0010  1 = Characters are outlined (djberger)
+        #                 5  0x0020  1 = Characters are shadowed (djberger)
+        #      4     2  Colour index (➜ 6.70)
+        #      6     2  Font weight (100-1000). Standard values are
+        #                           0x0190 (400) for normal text and
+        #                           0x02bc (700) for bold text.
+        #      8     2  Escapement type: 0x0000 = None
+        #                                0x0001 = Superscript
+        #                                0x0002 = Subscript
+        #     10     1  Underline type: 0x00 = None
+        #                               0x01 = Single
+        #                               0x02 = Double
+        #                               0x21 = Single accounting
+        #                               0x22 = Double accounting
+        #     11     1  Font family:
+        #               0x00 = None (unknown or don't care)
+        #               0x01 = Roman (variable width, serifed)
+        #               0x02 = Swiss (variable width, sans-serifed)
+        #               0x03 = Modern (fixed width, serifed or sans-serifed)
+        #               0x04 = Script (cursive)
+        #               0x05 = Decorative (specialised,
+        #                                  for example Old English, Fraktur)
+        #     12     1  Character set: 0x00 =   0 = ANSI Latin
+        #                              0x01 =   1 = System default
+        #                              0x02 =   2 = Symbol
+        #                              0x4d =  77 = Apple Roman
+        #                              0x80 = 128 = ANSI Japanese Shift-JIS
+        #                              0x81 = 129 = ANSI Korean (Hangul)
+        #                              0x82 = 130 = ANSI Korean (Johab)
+        #                              0x86 = 134 = ANSI Chinese Simplified GBK
+        #                              0x88 = 136 = ANSI Chinese Traditional BIG5
+        #                              0xa1 = 161 = ANSI Greek
+        #                              0xa2 = 162 = ANSI Turkish
+        #                              0xa3 = 163 = ANSI Vietnamese
+        #                              0xb1 = 177 = ANSI Hebrew
+        #                              0xb2 = 178 = ANSI Arabic
+        #                              0xba = 186 = ANSI Baltic
+        #                              0xcc = 204 = ANSI Cyrillic
+        #                              0xde = 222 = ANSI Thai
+        #                              0xee = 238 = ANSI Latin II (Central European)
+        #                              0xff = 255 = OEM Latin I
+        #     13     1  Not used
+        #     14  var.  Font name:
+        #               BIFF5/BIFF7: Byte string, 8-bit string length (➜ 3.3)
+        #               BIFF8: Unicode string, 8-bit string length (➜ 3.4)
+        name = client read_string(work[14..]), @workbook.encoding
+        font = Font.new name
+        size, opts, color, font.weight, escapement, underline,
+          family, encoding = work.unpack binfmt(:font)
+        font.size = size / TWIPS
+        font.italic = opts & 0x0002
+        font.strikeout = opts & 0x0008
+        font.outline = opts & 0x0010
+        font.shadow = opts & 0x0020
+        font.color = COLOR_CODES[color] || :text
+        font.escapement = ESCAPEMENT_TYPES[escapement]
+        font.underline = UNDERLINE_TYPES[underline]
+        font.family = FONT_FAMILIES[family]
+        font.encoding = FONT_ENCODINGS[encoding]
+        @workbook.add_font font
+      end
+
+      def read_format work, pos, len
+        # Offset  Size  Contents
+        #      0     2  Format index used in other records
+        #      2  var.  Number format string
+        #               (Unicode string, 16-bit string length, ➜ 3.4)
+        idx, = work.unpack "v"
+        value = read_string work[2..], 2
+        @formats.store idx, client(value, @workbook.encoding)
+      end
+
+      def read_formula worksheet, addr, work
+        # Offset  Size  Contents
+        #      0     2  Index to row
+        #      2     2  Index to column
+        #      4     2  Index to XF record (➜ 6.115)
+        #      6     8  Result of the formula. See below for details.
+        #     14     2  Option flags:
+        #               Bit  Mask    Contents
+        #                 0  0x0001  1 = Recalculate always
+        #                 1  0x0002  1 = Calculate on open
+        #                 3  0x0008  1 = Part of a shared formula
+        #     16     4  Not used
+        #     20  var.  Formula data (RPN token array, ➜ 4)
+        #               Offset  Size  Contents
+        #                    0     2  Size of the following formula data (sz)
+        #                    2    sz  Formula data (RPN token array)
+        #               [2+sz]  var.  (optional) Additional data for specific tokens
+        #                             (➜ 4.1.6, for example tArray token, ➜ 4.8.7)
+        #
+        # Result of the Formula
+        # Dependent on the type of value the formula returns, the result field has
+        # the following format:
+        #
+        # Result is a numeric value:
+        # Offset  Size  Contents
+        #      0     8  IEEE 754 floating-point value (64-bit double precision)
+        #
+        # Result is a string (the string follows in a STRING record, ➜ 6.98):
+        # Offset  Size  Contents
+        #      0     1  0x00 (identifier for a string value)
+        #      1     5  Not used
+        #      6     2  0xffff
+        # Note: In BIFF8 the string must not be empty. For empty cells there is a
+        # special identifier defined (see below).
+        #
+        # Result is a Boolean value:
+        # Offset  Size  Contents
+        #      0     1  0x01 (identifier for a Boolean value)
+        #      1     1  Not used
+        #      2     1  0 = FALSE, 1 = TRUE
+        #      3     3  Not used
+        #      6     2  0xffff
+        #
+        # Result is an error value:
+        # Offset  Size  Contents
+        #      0     1  0x02 (identifier for an error value)
+        #      1     1  Not used
+        #      2     1  Error code (➜ 3.7)
+        #      3     3  Not used
+        #      6     2  0xffff
+        #
+        # Result is an empty cell (BIFF8), for example an empty string:
+        # Offset  Size  Contents
+        #      0     1  0x03 (identifier for an empty cell)
+        #      1     5  Not used
+        #      6     2  0xffff
+        row, column, xf, rtype, rval, rcheck, opts = work.unpack "v3CxCx3v2"
+        formula = Formula.new
+        formula.shared = (opts & 0x08) > 0
+        formula.data = work[20..]
+        if rcheck != 0xffff || rtype > 3
+          value, = work.unpack "x6E"
+          unless value
+            # on architectures where sizeof(double) > 8
+            value, = work.unpack "x6e"
+          end
+          formula.value = value
+        elsif rtype == 0
+          pos, op, _len, work = get_next_chunk
+          if op == :sharedfmla
+            ## TODO: formula-support in 0.8.0
+            pos, op, _len, work = get_next_chunk
+          end
+          if op == :string
+            formula.value = client read_string(work, 2), @workbook.encoding
+          else
+            warn "String Value expected after Formula, but got #{op}"
+            formula.value = Error.new 0x2a
+            @pos = pos
+          end
+        elsif rtype == 1
+          formula.value = rval > 0
+        elsif rtype == 2
+          formula.value = Error.new rval
+        else
+          # leave the Formula value blank
+        end
+        set_cell worksheet, row, column, xf, formula
+      end
 
-  def read_workbook
-    previous_op = nil
-    while tuple = get_next_chunk
-      pos, op, len, work = tuple
-      case op
-      when @bof, :bof  # ●  BOF Type = worksheet (➜ 6.8)
-        return
-      when :eof        # ●  EOF ➜ 6.36
-        postread_workbook
-        return
-      when :datemode   # ○  DATEMODE ➜ 6.25
-        flag, _ = work.unpack 'v'
-        if flag == 1
-          @workbook.date_base = DateTime.new 1904, 1, 1
+      def read_hlink worksheet, work, pos_unused, len
+        # 6.53.1 Common Record Contents
+        # Offset  Size  Contents
+        #      0     8  Cell range address of all cells containing this hyperlink
+        #               (➜ 3.13.1)
+        #      8    16  GUID of StdLink:
+        #               D0 C9 EA 79 F9 BA CE 11 8C 82 00 AA 00 4B A9 0B
+        #               (79EAC9D0-BAF9-11CE-8C82-00AA004BA90B)
+        #     24     4  Unknown value: 0x00000002
+        #     28     4  Option flags (see below)
+        #               Bit  Mask        Contents
+        #                 0  0x00000001  0 = No link extant
+        #                                1 = File link or URL
+        #                 1  0x00000002  0 = Relative file path
+        #                                1 = Absolute path or URL
+        #           2 and 4  0x00000014  0 = No description
+        #                                1 (both bits) = Description
+        #                 3  0x00000008  0 = No text mark
+        #                                1 = Text mark
+        #                 7  0x00000080  0 = No target frame
+        #                                1 = Target frame
+        #                 8  0x00000100  0 = File link or URL
+        #                                1 = UNC path (incl. server name)
+        #--------------------------------------------------------------------------
+        #   [32]     4  (optional, see option flags) Character count of description
+        #               text, including trailing zero word (dl)
+        #   [36]  2∙dl  (optional, see option flags) Character array of description
+        #               text, no Unicode string header, always 16-bit characters,
+        #               zero-terminated
+        #--------------------------------------------------------------------------
+        # [var.]     4  (optional, see option flags) Character count of target
+        #               frame, including trailing zero word (fl)
+        # [var.]  2∙fl  (optional, see option flags) Character array of target
+        #               frame, no Unicode string header, always 16-bit characters,
+        #               zero-terminated
+        #--------------------------------------------------------------------------
+        #   var.  var.  Special data (➜ 6.53.2 and following)
+        #--------------------------------------------------------------------------
+        # [var.]     4  (optional, see option flags) Character count of the text
+        #               mark, including trailing zero word (tl)
+        # [var.]  2∙tl  (optional, see option flags) Character array of the text
+        #               mark without “#” sign, no Unicode string header, always
+        #               16-bit characters, zero-terminated
+        firstrow, lastrow, firstcol, lastcol, _, opts = work.unpack "v4H32x4V"
+        has_link = opts & 0x0001
+        desc = opts & 0x0014
+        textmark = opts & 0x0008
+        target = opts & 0x0080
+        unc = opts & 0x0100
+        link = Link.new
+        _, _ = nil
+        pos = 32
+        if desc > 0
+          description, pos = read_hlink_string work, pos
+          link << description
+        end
+        if target > 0
+          link.target_frame, pos = read_hlink_string work, pos
+        end
+        if unc > 0
+          # 6.53.4 Hyperlink to a File with UNC (Universal Naming Convention) Path
+          # These data fields are for UNC paths containing a server name (for
+          # instance “\\server\path\file.xls”). The lower 9 bits of the option
+          # flags field must be 1.x00x.xx112.
+          # Offset  Size  Contents
+          #      0     4  Character count of the UNC,
+          #               including trailing zero word (fl)
+          #      4  2∙fl  Character array of the UNC, no Unicode string header,
+          #               always 16-bit characters, zeroterminated.
+          link.url, pos = read_hlink_string work, pos
+        elsif has_link > 0
+          uid, = work.unpack "x#{pos}H32"
+          pos += 16
+          if uid == "e0c9ea79f9bace118c8200aa004ba90b"
+            # 6.53.2 Hyperlink containing a URL (Uniform Resource Locator)
+            # These data fields occur for links which are not local files or files
+            # in the local network (for instance HTTP and FTP links and e-mail
+            # addresses). The lower 9 bits of the option flags field must be
+            # 0.x00x.xx112 (x means optional, depending on hyperlink content). The
+            # GUID could be used to distinguish a URL from a file link.
+            # Offset  Size  Contents
+            #      0    16  GUID of URL Moniker:
+            #               E0 C9 EA 79 F9 BA CE 11 8C 82 00 AA 00 4B A9 0B
+            #               (79EAC9E0-BAF9-11CE-8C82-00AA004BA90B)
+            #     16     4  Size of character array of the URL, including trailing
+            #               zero word (us). There are us/2-1 characters in the
+            #               following string.
+            #     20    us  Character array of the URL, no Unicode string header,
+            #               always 16-bit characters, zeroterminated
+            size, = work.unpack "x#{pos}V"
+            pos += 4
+            data = work[pos, size].chomp "\000\000"
+            link.url = client data
+            pos += size
+          else
+            # 6.53.3 Hyperlink to a Local File
+            # These data fields are for links to files on local drives. The path of
+            # the file can be complete with drive letter (absolute) or relative to
+            # the location of the workbook. The lower 9 bits of the option flags
+            # field must be 0.x00x.xxx12. The GUID could be used to distinguish a
+            # URL from a file link.
+            # Offset  Size  Contents
+            #      0    16  GUID of File Moniker:
+            #               03 03 00 00 00 00 00 00 C0 00 00 00 00 00 00 46
+            #               (00000303-0000-0000-C000-000000000046)
+            #     16     2  Directory up-level count. Each leading “..\” in the
+            #               file link is deleted and increases this counter.
+            #     18     4  Character count of the shortened file path and name,
+            #               including trailing zero byte (sl)
+            #     22    sl  Character array of the shortened file path and name in
+            #               8.3-DOS-format. This field can be filled with a long
+            #               file name too. No Unicode string header, always 8-bit
+            #               characters, zeroterminated.
+            #  22+sl    24  Unknown byte sequence:
+            #               FF FF AD DE 00 00 00 00
+            #               00 00 00 00 00 00 00 00
+            #               00 00 00 00 00 00 00 00
+            #  46+sl     4  Size of the following file link field including string
+            #               length field and additional data field (sz). If sz is
+            #               zero, nothing will follow (except a text mark).
+            # [50+sl]    4  (optional) Size of character array of the extended file
+            #               path and name (xl). There are xl/2 characters in the
+            #               following string.
+            # [54+sl]    2  (optional) Unknown byte sequence: 03 00
+            # [56+sl]   xl  (optional) Character array of the extended file path
+            #               and name (xl), no Unicode string header, always 16-bit
+            #               characters, not zero-terminated
+            uplevel, count = work.unpack "x#{pos}vV"
+            pos += 6
+            # TODO: short file path may have any of the OEM encodings. Find out which
+            #       and use the #client method to convert the encoding.
+            prefix = internal("..\\", "UTF-8") * uplevel
+            link.dos = link.url = prefix << work[pos, count].chomp("\000")
+            pos += count + 24
+            total, size = work.unpack "x#{pos}V2"
+            pos += 10
+            if total > 0
+              link.url = client work[pos, size]
+              pos += size
+            end
+          end
         else
-          @workbook.date_base = DateTime.new 1899, 12, 31
+          # 6.53.5 Hyperlink to the Current Workbook
+          # In this case only the text mark field is present (optional with
+          # description).
+          # Example: The URL “#Sheet2!B1:C2” refers to the given range in the
+          # current workbook.
+          # The lower 9 bits of the option flags field must be 0.x00x.1x002.
+        end
+        if textmark > 0
+          link.fragment, _ = read_hlink_string work, pos
+        end
+        if link.empty?
+          link << link.href
         end
-      when :continue   # ○  CONTINUE ➜ 6.22
-        case previous_op
-        when :sst      # ●  SST ➜ 6.96
-          continue_sst work, pos, len
+        firstrow.upto lastrow do |row|
+          firstcol.upto lastcol do |col|
+            worksheet.add_link row, col, link
+          end
         end
-      when :codepage   # ○  CODEPAGE ➜ 6.17
-        read_codepage work, pos, len
-      when :boundsheet # ●● BOUNDSHEET ➜ 6.12
-        read_boundsheet work, pos, len
-      when :xf         # ●● XF ➜ 6.115
-        read_xf work, pos, len
-      when :sst        # ○  Shared String Table ➜ 5.11
-                       # ●  SST ➜ 6.96
-        read_sst work, pos, len
-        # TODO: implement memory-efficient sst handling, possibly in conjunction
-        #       with EXTSST
-      when :extsst     # ●  EXTSST ➜ 6.40
-        read_extsst work, pos, len
-      when :style      # ●● STYLE ➜ 6.99
-        read_style work, pos, len
-      when :format     # ○○ FORMAT (Number Format) ➜ 6.45
-        read_format work, pos, len
-      when :font       # ●● FONT ➜ 6.43
-        read_font work, pos, len
       end
-      previous_op = op unless op == :continue
-    end
-  end
-  def read_worksheet worksheet, offset
-    @pos = offset
-    @detected_rows = {}
-    @noteObjList = []
-    @noteList = []
-    @noteObject = nil
-    previous = nil
-    while tuple = get_next_chunk
-      pos, op, len, work = tuple
-      if((offset = @current_row_block_offset) && !in_row_block?(op, previous))
-        @current_row_block_offset = nil
-        offset[1] = pos - offset[0]
+
+      def read_hlink_string work, pos
+        count, = work.unpack "x#{pos}V"
+        len = count * 2
+        pos += 4
+        data = work[pos, len].chomp "\000\000"
+        pos += len
+        [client(data, "UTF-16LE"), pos]
       end
-      case op
-      when :eof        # ●  EOF ➜ 6.36
-        postread_worksheet worksheet
-        return
-      #when :uncalced   # ○  UNCALCED ➜ 6.104
-        # TODO: Formula support. Values were not calculated before saving
-        #warn <<-EOS
-        #  Some fields containig formulas were saved without a computed value.
-        #  Support Spreadsheet::Excel by implementing formula-calculations!
-        #EOS
-      #when :index      # ○  INDEX ➜ 5.7 (Row Blocks), ➜ 6.55
-        # TODO: if there are changes in rows, omit index when writing
-        #read_index worksheet, work, pos, len
-      when :guts       #    GUTS      5.53
-        read_guts worksheet, work, pos, len
-      when :colinfo    # ○○ COLINFO ➜ 6.18
-        read_colinfo worksheet, work, pos, len
-      when :dimensions # ●  DIMENSIONS ➜ 6.31
-        read_dimensions worksheet, work, pos, len
-      when :row        # ○○ Row Blocks ➜ 5.7
-                       # ●  ROW ➜ 6.83
-        set_row_address worksheet, work, pos, len
-      when :hlink
-        read_hlink worksheet, work, pos, len
-      when :window2
-        read_window2 worksheet, work, pos, len
-      when :mergedcells # ○○ MERGEDCELLS	➜ 5.67
-        read_merged_cells worksheet, work, pos, len
-      when :protect, :password
-        read_sheet_protection worksheet, op, work
-      when :note # a note references an :obj
-        read_note worksheet, work, pos, len
-      when :obj # it contains the author in the NTS structure
-        _ft, _cb, _ot, _objID = work.unpack('v4')
-        if _ot == 0x19
-          #puts "\nDEBUG: found Note Obj record"
-          @noteObject         = NoteObject.new
-          @noteObject.objID   = _objID
+
+      def read_index worksheet, work, pos, len
+        # Offset  Size  Contents
+        #      0     4  Not used
+        #      4     4  Index to first used row (rf, 0-based)
+        #      8     4  Index to first row of unused tail of sheet
+        #               (rl, last used row + 1, 0-based)
+        #     12     4  Absolute stream position of the
+        #               DEFCOLWIDTH record (➜ 6.29) of the current sheet. If this
+        #               record does not exist, the offset points to the record at
+        #               the position where the DEFCOLWIDTH record would occur.
+        #     16  4∙nm  Array of nm absolute stream positions to the
+        #               DBCELL record (➜ 6.26) of each Row Block
+        # TODO: use the index if it exists
+        # _, first_used, first_unused, defcolwidth, *indices = work.unpack 'V*'
+      end
+
+      def read_label worksheet, addr, work
+        # Offset  Size  Contents
+        #      0     2  Index to row
+        #      2     2  Index to column
+        #      4     2  Index to XF record (➜ 6.115)
+        #      6  var.  Unicode string, 16-bit string length (➜ 3.4)
+        row, column, xf = work.unpack "v3"
+        value = client read_string(work[6..], 2), @workbook.encoding
+        set_cell worksheet, row, column, xf, value
+      end
+
+      def read_labelsst worksheet, addr, work
+        # Offset  Size  Contents
+        #      0     2  Index to row
+        #      2     2  Index to column
+        #      4     2  Index to XF record (➜ 6.115)
+        #      6     4  Index into SST record (➜ 6.96)
+        row, column, xf, index = work.unpack binfmt(:labelsst)
+        set_cell worksheet, row, column, xf, worksheet.shared_string(index)
+      end
+
+      def read_mulblank worksheet, addr, work
+        # Offset  Size  Contents
+        #      0     2  Index to row
+        #      2     2  Index to first column (fc)
+        #      4  2∙nc  List of nc=lc-fc+1 16-bit indexes to XF records (➜ 6.115)
+        # 4+2∙nc     2  Index to last column (lc)
+        row, column, *xfs = work.unpack "v*"
+        xfs.pop #=> last_column
+        xfs.each_with_index { |xf, idx| set_cell worksheet, row, column + idx, xf }
+      end
+
+      def read_mulrk worksheet, addr, work
+        # Offset  Size  Contents
+        #      0     2  Index to row
+        #      2     2  Index to first column (fc)
+        #      4  6∙nc  List of nc=lc-fc+1 XF/RK structures. Each XF/RK contains:
+        #               Offset  Size  Contents
+        #                    0     2  Index to XF record (➜ 6.115)
+        #                    2     4  RK value (➜ 3.6)
+        # 4+6∙nc     2  Index to last column (lc)
+        row, column = work.unpack "v2"
+        4.step(work.size - 6, 6) do |idx|
+          xf, = work.unpack "x#{idx}v"
+          set_cell worksheet, row, column, xf, decode_rk(work[idx + 2, 4])
+          column += 1
         end
-        #p work
-      when :drawing # this can be followed by txo in case of a note
-        if previous == :obj
-          #puts "\nDEBUG: found MsDrawing record"
-          #p work
+      end
+
+      def read_number worksheet, addr, work
+        # Offset  Size  Contents
+        #      0     2  Index to row
+        #      2     2  Index to column
+        #      4     2  Index to XF record (➜ 6.115)
+        #      6     8  IEEE 754 floating-point value (64-bit double precision)
+        row, column, xf, value = work.unpack binfmt(:number)
+        set_cell worksheet, row, column, xf, value
+      end
+
+      def read_rk worksheet, addr, work
+        # Offset  Size  Contents
+        #      0     2  Index to row
+        #      2     2  Index to column
+        #      4     2  Index to XF record (➜ 6.115)
+        #      6     4  RK value (➜ 3.6)
+        row, column, xf = work.unpack "v3"
+        set_cell worksheet, row, column, xf, decode_rk(work[6, 4])
+      end
+
+      def read_row worksheet, addr
+        row = addr[:index]
+        @current_row_block.fetch [worksheet, row] do
+          @current_row_block.clear
+          cells = @current_row_block[[worksheet, row]] = Row.new(nil, row)
+          @pos = addr[:offset]
+          found = false
+          while (tuple = get_next_chunk)
+            pos, op, _, work = tuple
+            case op
+            when :eof      # ●  EOF ➜ 6.36 - we should only get here if there is just
+              #                 one Row-Block
+              @pos = pos
+              return cells
+            when :dbcell   # ○  DBCELL Stream offsets to the cell records of each row
+              return cells
+            when :row      # ○○ Row Blocks ➜ 5.7
+              # ●  ROW ➜ 6.83
+              # ignore, we already did these in read_worksheet
+              return cells if found
+            when :blank    # BLANK ➜ 6.7
+              found = true
+              read_blank worksheet, addr, work
+            when :boolerr  # BOOLERR ➜ 6.10
+              found = true
+              read_boolerr worksheet, addr, work
+            when 0x0002    # INTEGER ➜ 6.56 (BIFF2 only)
+              found = true
+              # TODO: implement for BIFF2 support
+            when :formula  # FORMULA ➜ 6.46
+              found = true
+              read_formula worksheet, addr, work
+            when :label    # LABEL ➜ 6.59 (BIFF2-BIFF7)
+              found = true
+              read_label worksheet, addr, work
+            when :labelsst # LABELSST ➜ 6.61 (BIFF8 only)
+              found = true
+              read_labelsst worksheet, addr, work
+            when :mulblank # MULBLANK ➜ 6.64 (BIFF5-BIFF8)
+              found = true
+              read_mulblank worksheet, addr, work
+            when :mulrk    # MULRK ➜ 6.65 (BIFF5-BIFF8)
+              found = true
+              read_mulrk worksheet, addr, work
+            when :number   # NUMBER ➜ 6.68
+              found = true
+              read_number worksheet, addr, work
+            when :rk       # RK ➜ 6.82 (BIFF3-BIFF8)
+              found = true
+              read_rk worksheet, addr, work
+            when :rstring  # RSTRING ➜ 6.84 (BIFF5/BIFF7)
+              found = true
+              read_rstring worksheet, addr, work
+            end
+          end
+          cells
         end
-      when :txo # this contains the length of the note text
-        if previous == :drawing
-          #puts "\nDEBUG: found TxO record"
-          #p work
+      end
+
+      def read_rstring worksheet, addr, work
+        # Offset  Size  Contents
+        #      0     2  Index to row
+        #      2     2  Index to column
+        #      4     2  Index to XF record (➜ 6.115)
+        #      6    sz  Unformatted Unicode string, 16-bit string length (➜ 3.4)
+        #   6+sz     2  Number of Rich-Text formatting runs (rt)
+        #   8+sz  4·rt  List of rt formatting runs (➜ 3.2)
+        row, column, xf = work.unpack "v3"
+        value = client read_string(work[6..], 2), @workbook.encoding
+        set_cell worksheet, row, column, xf, value
+      end
+
+      def read_window2 worksheet, work, pos, len
+        # This record contains additional settings for the document window
+        # (BIFF2-BIFF4) or for the window of a specific worksheet (BIFF5-BIFF8).
+        # It is part of the Sheet View Settings Block (➜ 4.5).
+        # Offset  Size  Contents
+        #      0     2  Option flags:
+        #               Bits  Mask    Contents
+        #                  0  0x0001  0 = Show formula results
+        #                             1 = Show formulas
+        #                  1  0x0002  0 = Do not show grid lines
+        #                             1 = Show grid lines
+        #                  2  0x0004  0 = Do not show sheet headers
+        #                             1 = Show sheet headers
+        #                  3  0x0008  0 = Panes are not frozen
+        #                             1 = Panes are frozen (freeze)
+        #                  4  0x0010  0 = Show zero values as empty cells
+        #                             1 = Show zero values
+        #                  5  0x0020  0 = Manual grid line colour
+        #                             1 = Automatic grid line colour
+        #                  6  0x0040  0 = Columns from left to right
+        #                             1 = Columns from right to left
+        #                  7  0x0080  0 = Do not show outline symbols
+        #                             1 = Show outline symbols
+        #                  8  0x0100  0 = Keep splits if pane freeze is removed
+        #                             1 = Remove splits if pane freeze is removed
+        #                  9  0x0200  0 = Sheet not selected
+        #                             1 = Sheet selected (BIFF5-BIFF8)
+        #                 10  0x0400  0 = Sheet not active
+        #                             1 = Sheet active (BIFF5-BIFF8)
+        #                 11  0x0800  0 = Show in normal view
+        #                             1 = Show in page break preview (BIFF8)
+        #      2     2  Index to first visible row
+        #      4     2  Index to first visible column
+        #      6     2  Colour index of grid line colour (➜ 5.74).
+        #               Note that in BIFF2-BIFF5 an RGB colour is written instead.
+        #      8     2  Not used
+        #     10     2  Cached magnification factor in page break preview (in percent)
+        #               0 = Default (60%)
+        #     12     2  Cached magnification factor in normal view (in percent)
+        #               0 = Default (100%)
+        #     14     4  Not used
+        flags, _ = work.unpack "v"
+        worksheet.selected = flags & 0x0200 > 0
+      end
+
+      def read_merged_cells worksheet, work, pos, len
+        # This record contains the addresses of merged cell ranges in the current sheet.
+        # Record MERGEDCELLS, BIFF8:
+        # Offset  Size  Contents
+        # 0	      var.	Cell range address list with merged ranges (➜ 2.5.15)
+        # If the record size exceeds the limit, it is not continued with a CONTINUE record,
+        # but another self-contained MERGEDCELLS record is started. The limit of 8224 bytes
+        # per record results in a maximum number of 1027 merged ranges.
+
+        worksheet.merged_cells.push(*read_range_address_list(work, len))
+        #
+        # A cell range address list consists of a field with the number of ranges and the list
+        # of the range addresses.
+        # Cell range address list, BIFF2-BIFF8:
+        # Offset  Size          Contents
+        # 0       2             Number of following cell range addresses (nm)
+        # 2	      6∙nm or 8∙nm  List of nm cell range addresses (➜ 2.5.14)
+        #
+      end
+
+      def read_workbook
+        previous_op = nil
+        while (tuple = get_next_chunk)
+          pos, op, len, work = tuple
+          case op
+          when @bof, :bof  # ●  BOF Type = worksheet (➜ 6.8)
+            return
+          when :eof        # ●  EOF ➜ 6.36
+            postread_workbook
+            return
+          when :datemode   # ○  DATEMODE ➜ 6.25
+            flag, _ = work.unpack "v"
+            @workbook.date_base = if flag == 1
+              DateTime.new 1904, 1, 1
+            else
+              DateTime.new 1899, 12, 31
+            end
+          when :continue   # ○  CONTINUE ➜ 6.22
+            case previous_op
+            when :sst      # ●  SST ➜ 6.96
+              continue_sst work, pos, len
+            end
+          when :codepage   # ○  CODEPAGE ➜ 6.17
+            read_codepage work, pos, len
+          when :boundsheet # ●● BOUNDSHEET ➜ 6.12
+            read_boundsheet work, pos, len
+          when :xf         # ●● XF ➜ 6.115
+            read_xf work, pos, len
+          when :sst        # ○  Shared String Table ➜ 5.11
+            # ●  SST ➜ 6.96
+            read_sst work, pos, len
+            # TODO: implement memory-efficient sst handling, possibly in conjunction
+            #       with EXTSST
+          when :extsst     # ●  EXTSST ➜ 6.40
+            read_extsst work, pos, len
+          when :style      # ●● STYLE ➜ 6.99
+            read_style work, pos, len
+          when :format     # ○○ FORMAT (Number Format) ➜ 6.45
+            read_format work, pos, len
+          when :font       # ●● FONT ➜ 6.43
+            read_font work, pos, len
+          end
+          previous_op = op unless op == :continue
         end
-      when :continue # this contains the actual note text
-        if previous == :txo && @noteObject
-          #puts "\nDEBUG: found Continue record"
-          continueFmt = work.unpack('C')
-          if (continueFmt.first == 0)
-             #puts "Picking compressed charset"
-             #Skip to offset due to 'v5C' used above
-             _text = work.unpack('@1C*')
-             @noteObject.text = _text.pack('C*')
-          elsif (continueFmt.first == 1)
-             #puts "Picking uncompressed charset"
-             _text = work.unpack('@1S*')
-             @noteObject.text = _text.pack('U*')
+      end
+
+      def read_worksheet worksheet, offset
+        @pos = offset
+        @detected_rows = {}
+        @note_ob_list = []
+        @note_list = []
+        @note_object = nil
+        previous = nil
+        while (tuple = get_next_chunk)
+          pos, op, len, work = tuple
+          if (offset = @current_row_block_offset) && !in_row_block?(op, previous)
+            @current_row_block_offset = nil
+            offset[1] = pos - offset[0]
           end
-          @noteObjList << @noteObject
+          case op
+          when :eof        # ●  EOF ➜ 6.36
+            postread_worksheet worksheet
+            return
+            # when :uncalced   # ○  UNCALCED ➜ 6.104
+            # TODO: Formula support. Values were not calculated before saving
+            # warn <<-EOS
+            #  Some fields containig formulas were saved without a computed value.
+            #  Support Spreadsheet::Excel by implementing formula-calculations!
+            # EOS
+            # when :index      # ○  INDEX ➜ 5.7 (Row Blocks), ➜ 6.55
+            # TODO: if there are changes in rows, omit index when writing
+            # read_index worksheet, work, pos, len
+          when :guts       #    GUTS      5.53
+            read_guts worksheet, work, pos, len
+          when :colinfo    # ○○ COLINFO ➜ 6.18
+            read_colinfo worksheet, work, pos, len
+          when :dimensions # ●  DIMENSIONS ➜ 6.31
+            read_dimensions worksheet, work, pos, len
+          when :row        # ○○ Row Blocks ➜ 5.7
+            # ●  ROW ➜ 6.83
+            set_row_address worksheet, work, pos, len
+          when :hlink
+            read_hlink worksheet, work, pos, len
+          when :window2
+            read_window2 worksheet, work, pos, len
+          when :mergedcells # ○○ MERGEDCELLS	➜ 5.67
+            read_merged_cells worksheet, work, pos, len
+          when :protect, :password
+            read_sheet_protection worksheet, op, work
+          when :note # a note references an :obj
+            read_note worksheet, work, pos, len
+          when :obj # it contains the author in the NTS structure
+            _ft, _cb, ot, obj_id = work.unpack("v4")
+            if ot == 0x19
+              # puts "\nDEBUG: found Note Obj record"
+              @note_object = NoteObject.new
+              @note_object.obj_id = obj_id
+            end
+            # p work
+          when :drawing # this can be followed by txo in case of a note
+            if previous == :obj
+              # puts "\nDEBUG: found MsDrawing record"
+              # p work
+            end
+          when :txo # this contains the length of the note text
+            if previous == :drawing
+              # puts "\nDEBUG: found TxO record"
+              # p work
+            end
+          when :continue # this contains the actual note text
+            if previous == :txo && @note_object
+              # puts "\nDEBUG: found Continue record"
+              continue_fmt = work.unpack("C")
+              if continue_fmt.first == 0
+                # puts "Picking compressed charset"
+                # Skip to offset due to 'v5C' used above
+                text = work.unpack("@1C*")
+                @note_object.text = text.pack("C*")
+              elsif continue_fmt.first == 1
+                # puts "Picking uncompressed charset"
+                text = work.unpack("@1S*")
+                @note_object.text = text.pack("U*")
+              end
+              @note_ob_list << @note_object
+            end
+          when :pagesetup
+            read_pagesetup(worksheet, work, pos, len)
+          when :leftmargin
+            worksheet.margins[:left] = work.unpack1(binfmt(:margin))
+          when :rightmargin
+            worksheet.margins[:right] = work.unpack1(binfmt(:margin))
+          when :topmargin
+            worksheet.margins[:top] = work.unpack1(binfmt(:margin))
+          when :bottommargin
+            worksheet.margins[:bottom] = work.unpack1(binfmt(:margin))
+          else
+            if ROW_BLOCK_OPS.include?(op)
+              set_missing_row_address worksheet, work, pos, len
+            end
+          end
+          previous = op
+          # previous = op unless op == :continue
         end
-      when :pagesetup
-        read_pagesetup(worksheet, work, pos, len)
-      when :leftmargin
-        worksheet.margins[:left] = work.unpack(binfmt(:margin))[0]
-      when :rightmargin
-        worksheet.margins[:right] = work.unpack(binfmt(:margin))[0]
-      when :topmargin
-        worksheet.margins[:top] = work.unpack(binfmt(:margin))[0]
-      when :bottommargin
-        worksheet.margins[:bottom] = work.unpack(binfmt(:margin))[0]
-      else
-        if ROW_BLOCK_OPS.include?(op)
-          set_missing_row_address worksheet, work, pos, len
+      end
+
+      def read_pagesetup(worksheet, work, pos, len)
+        worksheet.pagesetup.delete_if { true }
+        data = work.unpack(binfmt(:pagesetup))
+        worksheet.pagesetup[:orientation] = (data[5] == 0) ? :landscape : :portrait
+        worksheet.pagesetup[:adjust_to] = data[1]
+
+        worksheet.pagesetup[:orig_data] = data
+        # TODO: add options acording to specification
+      end
+
+      def read_guts worksheet, work, pos, len
+        # Offset Size Contents
+        #      0    2 Width of the area to display row outlines (left of the sheet), in pixel
+        #      2    2 Height of the area to display column outlines (above the sheet), in pixel
+        #      4    2 Number of visible row outline levels (used row levels + 1; or 0, if not used)
+        #      6    2 Number of visible column outline levels (used column levels + 1; or 0, if not used)
+        width, height, row_level, col_level = work.unpack "v4"
+        worksheet.guts[:width] = width
+        worksheet.guts[:height] = height
+        worksheet.guts[:row_level] = row_level
+        worksheet.guts[:col_level] = col_level
+      end
+
+      def read_style work, pos, len
+        # User-Defined Cell Styles:
+        # Offset  Size  Contents
+        #      0     2  Bit   Mask    Contents
+        #               11-0  0x0fff  Index to style XF record (➜ 6.115)
+        #                 15  0x8000  Always 0 for user-defined styles
+        #      2  var.  BIFF2-BIFF7:  Non-empty byte string,
+        #                             8-bit string length (➜ 3.3)
+        #               BIFF8:        Non-empty Unicode string,
+        #                             16-bit string length (➜ 3.4)
+        #
+        # Built-In Cell Styles
+        # Offset  Size  Contents
+        #      0     2  Bit   Mask    Contents
+        #               11-0  0x0FFF  Index to style XF record (➜ 6.115)
+        #                 15  0x8000  Always 1 for built-in styles
+        #      2     1  Identifier of the built-in cell style:
+        #               0x00 = Normal
+        #               0x01 = RowLevel_lv (see next field)
+        #               0x02 = ColLevel_lv (see next field)
+        #               0x03 = Comma
+        #               0x04 = Currency
+        #               0x05 = Percent
+        #               0x06 = Comma [0] (BIFF4-BIFF8)
+        #               0x07 = Currency [0] (BIFF4-BIFF8)
+        #               0x08 = Hyperlink (BIFF8)
+        #               0x09 = Followed Hyperlink (BIFF8)
+        #      3     1  Level for RowLevel or ColLevel style (zero-based, lv),
+        #               FFH otherwise
+        flags, = work.unpack "v"
+        xf_idx = flags & 0x0fff
+        xf = @workbook.format xf_idx
+        builtin = flags & 0x8000
+        if builtin == 0
+          xf.name = client read_string(work[2..], 2), @workbook.encoding
+        else
+          id, level = work.unpack "x2C2"
+          if (name = BUILTIN_STYLES[id])
+            name.sub "_lv", "_#{level}"
+            xf.name = client name, "UTF-8"
+          end
         end
       end
-      previous = op
-      #previous = op unless op == :continue
-    end
-  end
 
-  def read_pagesetup(worksheet, work, pos, len)
-    worksheet.pagesetup.delete_if { true }
-    data = work.unpack(binfmt(:pagesetup))
-    worksheet.pagesetup[:orientation] = data[5] == 0 ? :landscape : :portrait
-    worksheet.pagesetup[:adjust_to] = data[1]
+      def read_xf work, pos, len
+        # Offset  Size  Contents
+        #      0     2  Index to FONT record (➜ 6.43)
+        #      2     2  Index to FORMAT record (➜ 6.45)
+        #      4     2   Bit  Mask    Contents
+        #                2-0  0x0007  XF_TYPE_PROT – XF type, cell protection
+        #                             Bit  Mask  Contents
+        #                               0  0x01  1 = Cell is locked
+        #                               1  0x02  1 = Formula is hidden
+        #                               2  0x04  0 = Cell XF; 1 = Style XF
+        #               15-4  0xfff0  Index to parent style XF
+        #                             (always 0xfff in style XFs)
+        #      6     1   Bit  Mask    Contents
+        #                2-0  0x07    XF_HOR_ALIGN – Horizontal alignment
+        #                             Value  Horizontal alignment
+        #                             0x00   General
+        #                             0x01   Left
+        #                             0x02   Centred
+        #                             0x03   Right
+        #                             0x04   Filled
+        #                             0x05   Justified (BIFF4-BIFF8X)
+        #                             0x06   Centred across selection
+        #                                    (BIFF4-BIFF8X)
+        #                             0x07   Distributed (BIFF8X)
+        #                  3  0x08    1 = Text is wrapped at right border
+        #                6-4  0x70    XF_VERT_ALIGN – Vertical alignment
+        #                             Value  Vertical alignment
+        #                             0x00   Top
+        #                             0x01   Centred
+        #                             0x02   Bottom
+        #                             0x03   Justified (BIFF5-BIFF8X)
+        #                             0x04   Distributed (BIFF8X)
+        #      7     1  XF_ROTATION: Text rotation angle (see above)
+        #                Value  Text rotation
+        #                    0  Not rotated
+        #                 1-90  1 to 90 degrees counterclockwise
+        #               91-180  1 to 90 degrees clockwise
+        #                  255  Letters are stacked top-to-bottom,
+        #                       but not rotated
+        #      8     1   Bit  Mask    Contents
+        #                3-0  0x0f    Indent level
+        #                  4  0x10    1 = Shrink content to fit into cell
+        #                  5  0x40    1 = Merge Range (djberger)
+        #                7-6  0xc0    Text direction (BIFF8X only)
+        #                             0 = According to context
+        #                             1 = Left-to-right
+        #                             2 = Right-to-left
+        #      9     1   Bit  Mask    Contents
+        #                7-2  0xfc    XF_USED_ATTRIB – Used attributes
+        #                             Each bit describes the validity of a
+        #                             specific group of attributes. In cell XFs
+        #                             a cleared bit means the attributes of the
+        #                             parent style XF are used (but only if the
+        #                             attributes are valid there), a set bit
+        #                             means the attributes of this XF are used.
+        #                             In style XFs a cleared bit means the
+        #                             attribute setting is valid, a set bit
+        #                             means the attribute should be ignored.
+        #                             Bit  Mask  Contents
+        #                               0  0x01  Flag for number format
+        #                               1  0x02  Flag for font
+        #                               2  0x04  Flag for horizontal and
+        #                                        vertical alignment, text wrap,
+        #                                        indentation, orientation,
+        #                                        rotation, and text direction
+        #                               3  0x08  Flag for border lines
+        #                               4  0x10  Flag for background area style
+        #                               5  0x20  Flag for cell protection (cell
+        #                                        locked and formula hidden)
+        #     10     4  Cell border lines and background area:
+        #                 Bit  Mask        Contents
+        #                3- 0  0x0000000f  Left line style (➜ 3.10)
+        #                7- 4  0x000000f0  Right line style (➜ 3.10)
+        #               11- 8  0x00000f00  Top line style (➜ 3.10)
+        #               15-12  0x0000f000  Bottom line style (➜ 3.10)
+        #               22-16  0x007f0000  Colour index (➜ 6.70)
+        #                                  for left line colour
+        #               29-23  0x3f800000  Colour index (➜ 6.70)
+        #                                  for right line colour
+        #                  30  0x40000000  1 = Diagonal line
+        #                                  from top left to right bottom
+        #                  31  0x80000000  1 = Diagonal line
+        #                                  from bottom left to right top
+        #     14     4    Bit  Mask        Contents
+        #                6- 0  0x0000007f  Colour index (➜ 6.70)
+        #                                  for top line colour
+        #               13- 7  0x00003f80  Colour index (➜ 6.70)
+        #                                  for bottom line colour
+        #               20-14  0x001fc000  Colour index (➜ 6.70)
+        #                                  for diagonal line colour
+        #               24-21  0x01e00000  Diagonal line style (➜ 3.10)
+        #               31-26  0xfc000000  Fill pattern (➜ 3.11)
+        #     18     2    Bit  Mask        Contents
+        #                 6-0  0x007f      Colour index (➜ 6.70)
+        #                                  for pattern colour
+        #                13-7  0x3f80      Colour index (➜ 6.70)
+        #                                  for pattern background
+        fmt = Format.new
+        font_idx, numfmt, _, xf_align, xf_rotation, xf_indent, _,
+          xf_borders, xf_brdcolors, xf_pattern = work.unpack binfmt(:xf)
+        fmt.number_format = @formats[numfmt]
+        ## this appears to be undocumented: the first 4 fonts seem to be accessed
+        #  with a 0-based index, but all subsequent font indices are 1-based.
+        fmt.font = @workbook.font((font_idx > 3) ? font_idx - 1 : font_idx)
+        fmt.horizontal_align = NGILA_H_FX[xf_align & 0x07]
+        fmt.text_wrap = xf_align & 0x08 > 0
+        fmt.vertical_align = NGILA_V_FX[xf_align & 0x70]
+        fmt.rotation = if xf_rotation == 255
+          :stacked
+        elsif xf_rotation > 90
+          90 - xf_rotation
+        else
+          xf_rotation
+        end
+        fmt.indent_level = xf_indent & 0x0f
+        fmt.shrink = xf_indent & 0x10 > 0
+        fmt.text_direction = NOITCERID_TXET_FX[xf_indent & 0xc0]
+        fmt.left = XF_BORDER_LINE_STYLES[xf_borders & 0x0000000f]
+        fmt.right = XF_BORDER_LINE_STYLES[(xf_borders & 0x000000f0) >> 4]
+        fmt.top = XF_BORDER_LINE_STYLES[(xf_borders & 0x00000f00) >> 8]
+        fmt.bottom = XF_BORDER_LINE_STYLES[(xf_borders & 0x0000f000) >> 12]
+        fmt.left_color = COLOR_CODES[(xf_borders & 0x007f0000) >> 16] || :black
+        fmt.right_color = COLOR_CODES[(xf_borders & 0x3f800000) >> 23] || :black
+        fmt.cross_down = xf_borders & 0x40000000 > 0
+        fmt.cross_up = xf_borders & 0x80000000 > 0
+        if xf_brdcolors
+          fmt.top_color = COLOR_CODES[xf_brdcolors & 0x0000007f] || :black
+          fmt.bottom_color = COLOR_CODES[(xf_brdcolors & 0x00003f80) >> 7] || :black
+          fmt.diagonal_color = COLOR_CODES[(xf_brdcolors & 0x001fc000) >> 14] || :black
+          # fmt.diagonal_style = COLOR_CODES[xf_brdcolors & 0x01e00000]
+          fmt.pattern = (xf_brdcolors & 0xfc000000) >> 26
+        end
+        fmt.pattern_fg_color = COLOR_CODES[xf_pattern & 0x007f] || :border
+        fmt.pattern_bg_color = COLOR_CODES[(xf_pattern & 0x3f80) >> 7] || :pattern_bg
+        @workbook.add_format fmt
+      end
 
-    worksheet.pagesetup[:orig_data] = data
-    # TODO: add options acording to specification
-  end
+      def read_note worksheet, work, pos, len
+        # puts "\nDEBUG: found a note record in read_worksheet\n"
+        row, col, _, obj_id, obj_auth_en, obj_auth_len_fmt = work.unpack("v5C")
+        if obj_auth_en > 0
+          if obj_auth_len_fmt == 0
+            # puts "Picking compressed charset"
+            # Skip to offset due to 'v5C' used above
+            obj_auth = work.unpack("@11C" + (obj_auth_en - 1).to_s + "C")
+          elsif obj_auth_len_fmt == 1
+            # puts "Picking uncompressed charset"
+            obj_auth = work.unpack("@11S" + (obj_auth_en - 1).to_s + "S")
+          end
+          obj_auth = obj_auth.pack("C*")
+        else
+          obj_auth = ""
+        end
+        @note = Note.new
+        @note.length = len
+        @note.row = row
+        @note.col = col
+        @note.author = obj_auth
+        @note.obj_id = obj_id
+        # Pop it on the list to be sorted in postread_worksheet
+        @note_list << @note
+      end
 
-  def read_guts worksheet, work, pos, len
-    # Offset Size Contents
-    #      0    2 Width of the area to display row outlines (left of the sheet), in pixel
-    #      2    2 Height of the area to display column outlines (above the sheet), in pixel
-    #      4    2 Number of visible row outline levels (used row levels + 1; or 0, if not used)
-    #      6    2 Number of visible column outline levels (used column levels + 1; or 0, if not used)
-    width, height, row_level, col_level = work.unpack 'v4'
-    worksheet.guts[:width] = width
-    worksheet.guts[:height] = height
-    worksheet.guts[:row_level] = row_level
-    worksheet.guts[:col_level] = col_level
-  end
-  def read_style work, pos, len
-    # User-Defined Cell Styles:
-    # Offset  Size  Contents
-    #      0     2  Bit   Mask    Contents
-    #               11-0  0x0fff  Index to style XF record (➜ 6.115)
-    #                 15  0x8000  Always 0 for user-defined styles
-    #      2  var.  BIFF2-BIFF7:  Non-empty byte string,
-    #                             8-bit string length (➜ 3.3)
-    #               BIFF8:        Non-empty Unicode string,
-    #                             16-bit string length (➜ 3.4)
-    #
-    # Built-In Cell Styles
-    # Offset  Size  Contents
-    #      0     2  Bit   Mask    Contents
-    #               11-0  0x0FFF  Index to style XF record (➜ 6.115)
-    #                 15  0x8000  Always 1 for built-in styles
-    #      2     1  Identifier of the built-in cell style:
-    #               0x00 = Normal
-    #               0x01 = RowLevel_lv (see next field)
-    #               0x02 = ColLevel_lv (see next field)
-    #               0x03 = Comma
-    #               0x04 = Currency
-    #               0x05 = Percent
-    #               0x06 = Comma [0] (BIFF4-BIFF8)
-    #               0x07 = Currency [0] (BIFF4-BIFF8)
-    #               0x08 = Hyperlink (BIFF8)
-    #               0x09 = Followed Hyperlink (BIFF8)
-    #      3     1  Level for RowLevel or ColLevel style (zero-based, lv),
-    #               FFH otherwise
-    flags, = work.unpack 'v'
-    xf_idx = flags & 0x0fff
-    xf = @workbook.format xf_idx
-    builtin = flags & 0x8000
-    if builtin == 0
-      xf.name = client read_string(work[2..-1], 2), @workbook.encoding
-    else
-      id, level = work.unpack 'x2C2'
-      if name = BUILTIN_STYLES[id]
-        name.sub '_lv', "_#{level.to_s}"
-        xf.name = client name, 'UTF-8'
+      def read_sheet_protection worksheet, op, data
+        case op
+        when :protect
+          worksheet.protect! if data.unpack1("v") == 1
+        when :password
+          worksheet.password_hash = data.unpack1("v")
+        end
+      end
+
+      def set_cell worksheet, row, column, xf, value = nil
+        cells = @current_row_block[[worksheet, row]] ||= Row.new(nil, row)
+        cells.formats[column] = @workbook.format(xf) unless xf == 0
+        cells[column] = value
+      end
+
+      def set_missing_row_address worksheet, work, pos, len
+        # Offset  Size  Contents
+        #      0     2  Index of this row
+        #      2     2  Index to this column
+        row_index, _ = work.unpack "v2"
+        unless worksheet.offsets[row_index]
+          @current_row_block_offset ||= [pos]
+          data = {
+            index: row_index,
+            row_block: @current_row_block_offset,
+            offset: @current_row_block_offset[0]
+          }
+          worksheet.set_row_address row_index, data
+        end
+      end
+
+      def set_row_address worksheet, work, pos, len
+        # Offset  Size  Contents
+        #      0     2  Index of this row
+        #      2     2  Index to column of the first cell which
+        #               is described by a cell record
+        #      4     2  Index to column of the last cell which is
+        #               described by a cell record, increased by 1
+        #      6     2  Bit   Mask    Contents
+        #               14-0  0x7fff  Height of the row, in twips = 1/20 of a point
+        #                 15  0x8000  0 = Row has custom height;
+        #                             1 = Row has default height
+        #      8     2  Not used
+        #     10     2  In BIFF3-BIFF4 this field contains a relative offset to
+        #               calculate stream position of the first cell record for this
+        #               row (➜ 5.7.1). In BIFF5-BIFF8 this field is not used
+        #               anymore, but the DBCELL record (➜ 6.26) instead.
+        #     12     4  Option flags and default row formatting:
+        #                  Bit  Mask        Contents
+        #                  2-0  0x00000007  Outline level of the row
+        #                    4  0x00000010  1 = Outline group starts or ends here
+        #                                       (depending on where the outline
+        #                                       buttons are located, see WSBOOL
+        #                                       record, ➜ 6.113), and is collapsed
+        #                    5  0x00000020  1 = Row is hidden (manually, or by a
+        #                                       filter or outline group)
+        #                    6  0x00000040  1 = Row height and default font height
+        #                                       do not match
+        #                    7  0x00000080  1 = Row has explicit default format (fl)
+        #                    8  0x00000100  Always 1
+        #                27-16  0x0fff0000  If fl = 1: Index to default XF record
+        #                                              (➜ 6.115)
+        #                   28  0x10000000  1 = Additional space above the row.
+        #                                       This flag is set, if the upper
+        #                                       border of at least one cell in this
+        #                                       row or if the lower border of at
+        #                                       least one cell in the row above is
+        #                                       formatted with a thick line style.
+        #                                       Thin and medium line styles are not
+        #                                       taken into account.
+        #                   29  0x20000000  1 = Additional space below the row.
+        #                                       This flag is set, if the lower
+        #                                       border of at least one cell in this
+        #                                       row or if the upper border of at
+        #                                       least one cell in the row below is
+        #                                       formatted with a medium or thick
+        #                                       line style. Thin line styles are
+        #                                       not taken into account.
+        @current_row_block_offset ||= [pos]
+        index, first_used, first_unused, height, flags = work.unpack binfmt(:row)
+        height &= 0x7fff
+        format = nil
+        # TODO: read attributes from work[13,3], read flags
+        attrs = {
+          default_format: format,
+          first_used: first_used,
+          first_unused: first_unused,
+          index: index,
+          row_block: @current_row_block_offset,
+          offset: @current_row_block_offset[0],
+          outline_level: flags & 0x00000007,
+          collapsed: (flags & 0x0000010) > 0,
+          hidden: (flags & 0x0000020) > 0
+        }
+        if (flags & 0x00000040) > 0
+          attrs.store :height, height / TWIPS
+        end
+        if (flags & 0x00000080) > 0
+          xf = (flags & 0x0fff0000) >> 16
+          attrs.store :default_format, @workbook.format(xf)
+        end
+        # TODO: Row spacing
+        worksheet.set_row_address index, attrs
+      end
+
+      def setup io
+        ## Reading from StringIO fails without forced encoding
+        if Gem::Version.new(RUBY_VERSION) >= Gem::Version.new("2.3.0")
+          io.set_encoding("ASCII-8BIT")
+        elsif io.respond_to?(:string) && (str = io.string) && str.respond_to?(:force_encoding)
+          str.force_encoding("ASCII-8BIT")
+        end
+        io.rewind
+        @ole = Ole::Storage.open io
+        @workbook = Workbook.new io, {}
+        %w[Book Workbook BOOK WORKBOOK book workbook].any? do |name|
+          @book = begin
+            @ole.file.open(name)
+          rescue
+            false
+          end
+        end
+        raise "could not locate a workbook, possibly an empty file passed" unless @book
+        @data = @book.read
+        read_bof
+        @workbook.ole = @book
+        @workbook.bof = @bof
+        @workbook.version = @version
+        biff = @workbook.biff_version
+        extend_reader biff
+        extend_internals biff
+      end
+
+      private
+
+      def extend_internals version
+        require "spreadsheet/excel/internals/biff%i" % version
+        extend Internals.const_get("Biff%i" % version)
+        ## spreadsheets may not include a codepage record.
+        @workbook.encoding = encoding 850 if version < 8
+      rescue LoadError
+      end
+
+      def extend_reader version
+        require "spreadsheet/excel/reader/biff%i" % version
+        extend Reader.const_get("Biff%i" % version)
+      rescue LoadError
       end
     end
   end
-  def read_xf work, pos, len
-    # Offset  Size  Contents
-    #      0     2  Index to FONT record (➜ 6.43)
-    #      2     2  Index to FORMAT record (➜ 6.45)
-    #      4     2   Bit  Mask    Contents
-    #                2-0  0x0007  XF_TYPE_PROT – XF type, cell protection
-    #                             Bit  Mask  Contents
-    #                               0  0x01  1 = Cell is locked
-    #                               1  0x02  1 = Formula is hidden
-    #                               2  0x04  0 = Cell XF; 1 = Style XF
-    #               15-4  0xfff0  Index to parent style XF
-    #                             (always 0xfff in style XFs)
-    #      6     1   Bit  Mask    Contents
-    #                2-0  0x07    XF_HOR_ALIGN – Horizontal alignment
-    #                             Value  Horizontal alignment
-    #                             0x00   General
-    #                             0x01   Left
-    #                             0x02   Centred
-    #                             0x03   Right
-    #                             0x04   Filled
-    #                             0x05   Justified (BIFF4-BIFF8X)
-    #                             0x06   Centred across selection
-    #                                    (BIFF4-BIFF8X)
-    #                             0x07   Distributed (BIFF8X)
-    #                  3  0x08    1 = Text is wrapped at right border
-    #                6-4  0x70    XF_VERT_ALIGN – Vertical alignment
-    #                             Value  Vertical alignment
-    #                             0x00   Top
-    #                             0x01   Centred
-    #                             0x02   Bottom
-    #                             0x03   Justified (BIFF5-BIFF8X)
-    #                             0x04   Distributed (BIFF8X)
-    #      7     1  XF_ROTATION: Text rotation angle (see above)
-    #                Value  Text rotation
-    #                    0  Not rotated
-    #                 1-90  1 to 90 degrees counterclockwise
-    #               91-180  1 to 90 degrees clockwise
-    #                  255  Letters are stacked top-to-bottom,
-    #                       but not rotated
-    #      8     1   Bit  Mask    Contents
-    #                3-0  0x0f    Indent level
-    #                  4  0x10    1 = Shrink content to fit into cell
-    #                  5  0x40    1 = Merge Range (djberger)
-    #                7-6  0xc0    Text direction (BIFF8X only)
-    #                             0 = According to context
-    #                             1 = Left-to-right
-    #                             2 = Right-to-left
-    #      9     1   Bit  Mask    Contents
-    #                7-2  0xfc    XF_USED_ATTRIB – Used attributes
-    #                             Each bit describes the validity of a
-    #                             specific group of attributes. In cell XFs
-    #                             a cleared bit means the attributes of the
-    #                             parent style XF are used (but only if the
-    #                             attributes are valid there), a set bit
-    #                             means the attributes of this XF are used.
-    #                             In style XFs a cleared bit means the
-    #                             attribute setting is valid, a set bit
-    #                             means the attribute should be ignored.
-    #                             Bit  Mask  Contents
-    #                               0  0x01  Flag for number format
-    #                               1  0x02  Flag for font
-    #                               2  0x04  Flag for horizontal and
-    #                                        vertical alignment, text wrap,
-    #                                        indentation, orientation,
-    #                                        rotation, and text direction
-    #                               3  0x08  Flag for border lines
-    #                               4  0x10  Flag for background area style
-    #                               5  0x20  Flag for cell protection (cell
-    #                                        locked and formula hidden)
-    #     10     4  Cell border lines and background area:
-    #                 Bit  Mask        Contents
-    #                3- 0  0x0000000f  Left line style (➜ 3.10)
-    #                7- 4  0x000000f0  Right line style (➜ 3.10)
-    #               11- 8  0x00000f00  Top line style (➜ 3.10)
-    #               15-12  0x0000f000  Bottom line style (➜ 3.10)
-    #               22-16  0x007f0000  Colour index (➜ 6.70)
-    #                                  for left line colour
-    #               29-23  0x3f800000  Colour index (➜ 6.70)
-    #                                  for right line colour
-    #                  30  0x40000000  1 = Diagonal line
-    #                                  from top left to right bottom
-    #                  31  0x80000000  1 = Diagonal line
-    #                                  from bottom left to right top
-    #     14     4    Bit  Mask        Contents
-    #                6- 0  0x0000007f  Colour index (➜ 6.70)
-    #                                  for top line colour
-    #               13- 7  0x00003f80  Colour index (➜ 6.70)
-    #                                  for bottom line colour
-    #               20-14  0x001fc000  Colour index (➜ 6.70)
-    #                                  for diagonal line colour
-    #               24-21  0x01e00000  Diagonal line style (➜ 3.10)
-    #               31-26  0xfc000000  Fill pattern (➜ 3.11)
-    #     18     2    Bit  Mask        Contents
-    #                 6-0  0x007f      Colour index (➜ 6.70)
-    #                                  for pattern colour
-    #                13-7  0x3f80      Colour index (➜ 6.70)
-    #                                  for pattern background
-    fmt = Format.new
-    font_idx, numfmt, _, xf_align, xf_rotation, xf_indent, _,
-      xf_borders, xf_brdcolors, xf_pattern = work.unpack binfmt(:xf)
-    fmt.number_format = @formats[numfmt]
-    ## this appears to be undocumented: the first 4 fonts seem to be accessed
-    #  with a 0-based index, but all subsequent font indices are 1-based.
-    fmt.font = @workbook.font(font_idx > 3 ? font_idx - 1 : font_idx)
-    fmt.horizontal_align = NGILA_H_FX[xf_align & 0x07]
-    fmt.text_wrap = xf_align & 0x08 > 0
-    fmt.vertical_align = NGILA_V_FX[xf_align & 0x70]
-    fmt.rotation = if xf_rotation == 255
-                     :stacked
-                   elsif xf_rotation > 90
-                     90 - xf_rotation
-                   else
-                     xf_rotation
-                   end
-    fmt.indent_level = xf_indent & 0x0f
-    fmt.shrink = xf_indent & 0x10 > 0
-    fmt.text_direction = NOITCERID_TXET_FX[xf_indent & 0xc0]
-    fmt.left           = XF_BORDER_LINE_STYLES[xf_borders & 0x0000000f]
-    fmt.right          = XF_BORDER_LINE_STYLES[(xf_borders & 0x000000f0) >>  4]
-    fmt.top            = XF_BORDER_LINE_STYLES[(xf_borders & 0x00000f00) >>  8]
-    fmt.bottom         = XF_BORDER_LINE_STYLES[(xf_borders & 0x0000f000) >> 12]
-    fmt.left_color     = COLOR_CODES[(xf_borders & 0x007f0000) >> 16] || :black
-    fmt.right_color    = COLOR_CODES[(xf_borders & 0x3f800000) >> 23] || :black
-    fmt.cross_down     = xf_borders & 0x40000000 > 0
-    fmt.cross_up       = xf_borders & 0x80000000 > 0
-		if xf_brdcolors
-    	fmt.top_color      = COLOR_CODES[xf_brdcolors & 0x0000007f] || :black
-    	fmt.bottom_color   = COLOR_CODES[(xf_brdcolors & 0x00003f80) >> 7] || :black
-    	fmt.diagonal_color = COLOR_CODES[(xf_brdcolors & 0x001fc000) >> 14] || :black
-    	#fmt.diagonal_style = COLOR_CODES[xf_brdcolors & 0x01e00000]
-    	fmt.pattern        = (xf_brdcolors & 0xfc000000) >> 26
-		end
-    fmt.pattern_fg_color = COLOR_CODES[xf_pattern & 0x007f] || :border
-    fmt.pattern_bg_color = COLOR_CODES[(xf_pattern & 0x3f80) >> 7] || :pattern_bg
-    @workbook.add_format fmt
-  end
-  def read_note worksheet, work, pos, len
-    #puts "\nDEBUG: found a note record in read_worksheet\n"
-    row, col, _, _objID, _objAuthLen, _objAuthLenFmt = work.unpack('v5C')
-    if (_objAuthLen > 0)
-       if (_objAuthLenFmt == 0)
-          #puts "Picking compressed charset"
-          #Skip to offset due to 'v5C' used above
-          _objAuth = work.unpack('@11C' + (_objAuthLen-1).to_s + 'C')
-       elsif (_objAuthLenFmt == 1)
-          #puts "Picking uncompressed charset"
-          _objAuth = work.unpack('@11S' + (_objAuthLen-1).to_s + 'S')
-       end
-       _objAuth = _objAuth.pack('C*')
-    else
-       _objAuth = ""
-    end
-    @note = Note.new
-    @note.length = len
-    @note.row    = row
-    @note.col    = col
-    @note.author = _objAuth
-    @note.objID  = _objID
-    #Pop it on the list to be sorted in postread_worksheet
-    @noteList << @note
-  end
-  def read_sheet_protection worksheet, op, data
-    case op
-    when :protect
-      worksheet.protect! if data.unpack('v').first == 1
-    when :password
-      worksheet.password_hash = data.unpack('v').first
-    end
-  end
-  def set_cell worksheet, row, column, xf, value=nil
-    cells = @current_row_block[[worksheet, row]] ||= Row.new(nil, row)
-    cells.formats[column] = @workbook.format(xf) unless xf == 0
-    cells[column] = value
-  end
-  def set_missing_row_address worksheet, work, pos, len
-    # Offset  Size  Contents
-    #      0     2  Index of this row
-    #      2     2  Index to this column
-    row_index, _ = work.unpack 'v2'
-    unless worksheet.offsets[row_index]
-      @current_row_block_offset ||= [pos]
-      data = {
-        :index          => row_index,
-        :row_block      => @current_row_block_offset,
-        :offset         => @current_row_block_offset[0],
-      }
-      worksheet.set_row_address row_index, data
-    end
-  end
-  def set_row_address worksheet, work, pos, len
-    # Offset  Size  Contents
-    #      0     2  Index of this row
-    #      2     2  Index to column of the first cell which
-    #               is described by a cell record
-    #      4     2  Index to column of the last cell which is
-    #               described by a cell record, increased by 1
-    #      6     2  Bit   Mask    Contents
-    #               14-0  0x7fff  Height of the row, in twips = 1/20 of a point
-    #                 15  0x8000  0 = Row has custom height;
-    #                             1 = Row has default height
-    #      8     2  Not used
-    #     10     2  In BIFF3-BIFF4 this field contains a relative offset to
-    #               calculate stream position of the first cell record for this
-    #               row (➜ 5.7.1). In BIFF5-BIFF8 this field is not used
-    #               anymore, but the DBCELL record (➜ 6.26) instead.
-    #     12     4  Option flags and default row formatting:
-    #                  Bit  Mask        Contents
-    #                  2-0  0x00000007  Outline level of the row
-    #                    4  0x00000010  1 = Outline group starts or ends here
-    #                                       (depending on where the outline
-    #                                       buttons are located, see WSBOOL
-    #                                       record, ➜ 6.113), and is collapsed
-    #                    5  0x00000020  1 = Row is hidden (manually, or by a
-    #                                       filter or outline group)
-    #                    6  0x00000040  1 = Row height and default font height
-    #                                       do not match
-    #                    7  0x00000080  1 = Row has explicit default format (fl)
-    #                    8  0x00000100  Always 1
-    #                27-16  0x0fff0000  If fl = 1: Index to default XF record
-    #                                              (➜ 6.115)
-    #                   28  0x10000000  1 = Additional space above the row.
-    #                                       This flag is set, if the upper
-    #                                       border of at least one cell in this
-    #                                       row or if the lower border of at
-    #                                       least one cell in the row above is
-    #                                       formatted with a thick line style.
-    #                                       Thin and medium line styles are not
-    #                                       taken into account.
-    #                   29  0x20000000  1 = Additional space below the row.
-    #                                       This flag is set, if the lower
-    #                                       border of at least one cell in this
-    #                                       row or if the upper border of at
-    #                                       least one cell in the row below is
-    #                                       formatted with a medium or thick
-    #                                       line style. Thin line styles are
-    #                                       not taken into account.
-    @current_row_block_offset ||= [pos]
-    index, first_used, first_unused, height, flags = work.unpack binfmt(:row)
-    height &= 0x7fff
-    format = nil
-    # TODO: read attributes from work[13,3], read flags
-    attrs = {
-      :default_format => format,
-      :first_used     => first_used,
-      :first_unused   => first_unused,
-      :index          => index,
-      :row_block      => @current_row_block_offset,
-      :offset         => @current_row_block_offset[0],
-      :outline_level  => flags & 0x00000007,
-      :collapsed      => (flags & 0x0000010) > 0,
-      :hidden         => (flags & 0x0000020) > 0,
-    }
-    if (flags & 0x00000040) > 0
-      attrs.store :height, height / TWIPS
-    end
-    if (flags & 0x00000080) > 0
-      xf = (flags & 0x0fff0000) >> 16
-      attrs.store :default_format, @workbook.format(xf)
-    end
-    # TODO: Row spacing
-    worksheet.set_row_address index, attrs
-  end
-  def setup io
-    ## Reading from StringIO fails without forced encoding
-    if Gem::Version.new(RUBY_VERSION) >= Gem::Version.new('2.3.0')
-      io.set_encoding('ASCII-8BIT')
-    elsif io.respond_to?(:string) && (str = io.string) && str.respond_to?(:force_encoding)
-      str.force_encoding('ASCII-8BIT')
-    end
-    io.rewind
-    @ole = Ole::Storage.open io
-    @workbook = Workbook.new io, {}
-    %w{Book Workbook BOOK WORKBOOK book workbook}.any? do |name|
-      @book = @ole.file.open(name) rescue false
-    end
-    raise RuntimeError, "could not locate a workbook, possibly an empty file passed" unless @book
-    @data = @book.read
-    read_bof
-    @workbook.ole = @book
-    @workbook.bof = @bof
-    @workbook.version = @version
-    biff = @workbook.biff_version
-    extend_reader biff
-    extend_internals biff
-  end
-  private
-  def extend_internals version
-    require 'spreadsheet/excel/internals/biff%i' % version
-    extend Internals.const_get('Biff%i' % version)
-    ## spreadsheets may not include a codepage record.
-    @workbook.encoding = encoding 850 if version < 8
-  rescue LoadError
-  end
-  def extend_reader version
-    require 'spreadsheet/excel/reader/biff%i' % version
-    extend Reader.const_get('Biff%i' % version)
-  rescue LoadError
-  end
-end
-  end
 end