omnizip 0.4.7-aarch64-linux

data/lib/omnizip/formats/rar/rar5/compression/lzss.rb

@@ -0,0 +1,584 @@
+# frozen_string_literal: true
+
+require "stringio"
+
+module Omnizip
+  module Formats
+    module Rar
+      module Rar5
+        module Compression
+          # RAR5 LZSS compression method
+          #
+          # RAR5 compression methods 1-5 use a proprietary LZSS-based algorithm
+          # with Huffman coding. This is the algorithm used by official RAR tools.
+          #
+          # Based on libarchive/archive_read_support_format_rar5.c
+          #
+          class Lzss
+            # Compression method identifiers
+            METHOD_STORE = 0
+            METHOD_FASTEST = 1
+            METHOD_FAST = 2
+            METHOD_NORMAL = 3
+            METHOD_GOOD = 4
+            METHOD_BEST = 5
+
+            # Huffman code constants (from libarchive)
+            HUFF_BC = 20    # Number of bit length codes
+            HUFF_NC = 306   # Number of literal/length codes
+            HUFF_DC = 64    # Number of distance codes
+            HUFF_LDC = 16   # Number of low distance codes
+            HUFF_RC = 44    # Number of repeat codes
+            HUFF_TABLE_SIZE = HUFF_NC + HUFF_DC + HUFF_LDC + HUFF_RC
+
+            # Distance cache size
+            DIST_CACHE_SIZE = 4
+
+            # Minimum match length
+            MIN_MATCH = 3
+
+            class << self
+              # Check if LZSS compression is available for official RAR compatibility
+              #
+              # @return [Boolean] true if implemented
+              def available?
+                # Decoder is implemented; the encoder is NOT compatible
+                # with official RAR tools (unrar "tests" its output as
+                # OK while extracting empty files). Callers fall back
+                # to STORE until a RAR-compatible encoder lands.
+                false
+              end
+
+              # Compress data using RAR5 LZSS
+              #
+              # @param data [String] Data to compress
+              # @param options [Hash] Compression options
+              # @option options [Integer] :level Compression level (1-5)
+              # @option options [Integer] :dict_size Dictionary size
+              # @return [Hash] Hash with :data, :properties, and :method
+              def compress(data, options = {})
+                level = options[:level] || METHOD_NORMAL
+                options[:dict_size] || dictionary_size_for_level(level)
+
+                # For now, use STORE method until encoder is compatible
+                # with official RAR tools
+                {
+                  data: data,
+                  properties: nil,
+                  method: METHOD_STORE,
+                }
+              end
+
+              # Decompress RAR5 LZSS data
+              #
+              # @param data [String] Compressed data
+              # @param options [Hash] Decompression options
+              # @option options [Integer] :uncompressed_size Expected size
+              # @option options [Integer] :window_size Dictionary size
+              # @return [String] Decompressed data
+              def decompress(data, options = {})
+                uncompressed_size = options[:uncompressed_size]
+                window_size = options[:window_size] || (1 << 20) # Default 1MB
+
+                decoder = Decoder.new(data, window_size)
+                decoder.decode(uncompressed_size)
+              end
+
+              # Get compression method identifier
+              #
+              # @param level [Integer] Compression level (1-5)
+              # @return [Integer] Method ID
+              def method_id(level = METHOD_NORMAL)
+                level.clamp(METHOD_FASTEST, METHOD_BEST)
+              end
+
+              # Get compression info VINT value
+              #
+              # @param level [Integer] Compression level (1-5)
+              # @return [Integer] Compression info value
+              def compression_info(level = METHOD_NORMAL)
+                method = method_id(level)
+                method & 0x3F
+              end
+
+              private
+
+              # Get dictionary size for compression level
+              def dictionary_size_for_level(level)
+                1 << case level
+                     when 1 then 18   # 256 KB
+                     when 2 then 20   # 1 MB
+                     when 3 then 22   # 4 MB
+                     when 4 then 23   # 8 MB
+                     when 5 then 24   # 16 MB
+                     else 22 # 4 MB default
+                     end
+              end
+            end
+
+            # Bit reader for reading individual bits from compressed data
+            #
+            class BitReader
+              def initialize(data)
+                @data = data
+                @byte_pos = 0
+                @bit_pos = 0
+              end
+
+              # Read up to 16 bits
+              def read_bits(num_bits)
+                return 0 if num_bits.zero?
+
+                result = 0
+                bits_read = 0
+
+                while bits_read < num_bits
+                  return nil if @byte_pos >= @data.bytesize
+
+                  byte = @data.getbyte(@byte_pos)
+                  bits_available = 8 - @bit_pos
+                  bits_needed = num_bits - bits_read
+                  bits_to_read = [bits_available, bits_needed].min
+
+                  mask = ((1 << bits_to_read) - 1) << @bit_pos
+                  bits = (byte & mask) >> @bit_pos
+
+                  result |= bits << bits_read
+                  bits_read += bits_to_read
+                  @bit_pos += bits_to_read
+
+                  if @bit_pos >= 8
+                    @bit_pos = 0
+                    @byte_pos += 1
+                  end
+                end
+
+                result
+              end
+
+              # Skip specified number of bits
+              def skip_bits(num_bits)
+                @bit_pos += num_bits
+                while @bit_pos >= 8
+                  @bit_pos -= 8
+                  @byte_pos += 1
+                end
+              end
+
+              # Read 32 bits
+              def read_bits_32(num_bits)
+                return 0 if num_bits.zero?
+
+                result = 0
+                bits_read = 0
+
+                while bits_read < num_bits && @byte_pos < @data.bytesize
+                  byte = @data.getbyte(@byte_pos)
+                  bits_available = 8 - @bit_pos
+                  bits_needed = num_bits - bits_read
+                  bits_to_read = [bits_available, bits_needed].min
+
+                  mask = ((1 << bits_to_read) - 1) << @bit_pos
+                  bits = (byte & mask) >> @bit_pos
+
+                  result |= bits << bits_read
+                  bits_read += bits_to_read
+                  @bit_pos += bits_to_read
+
+                  if @bit_pos >= 8
+                    @bit_pos = 0
+                    @byte_pos += 1
+                  end
+                end
+
+                result
+              end
+
+              def end_of_data?(block_size)
+                @byte_pos >= block_size || (@byte_pos == block_size - 1 && @bit_pos >= 8)
+              end
+
+              attr_accessor :byte_pos, :bit_pos
+            end
+
+            # Huffman decode table
+            #
+            class HuffmanTable
+              attr_reader :size, :decode_len, :decode_pos, :decode_num
+
+              def initialize(size)
+                @size = size
+                @decode_len = Array.new(16, 0)
+                @decode_pos = Array.new(16, 0)
+                @decode_num = Array.new(size, 0)
+                @quick_bits = 0
+                @quick_len = Array.new(65536, 0)
+                @quick_num = Array.new(65536, 0)
+              end
+
+              # Build decode tables from bit lengths
+              # Based on libarchive's create_decode_tables()
+              def build(bit_lengths)
+                # Count codes for each bit length
+                len_count = Array.new(16, 0)
+                bit_lengths.each do |len|
+                  len_count[len] += 1 if len.positive? && len < 16
+                end
+
+                # Calculate decode_len and decode_pos
+                @decode_pos[0] = 0
+                @decode_len[0] = 0
+
+                upper_limit = 0
+                (1..15).each do |i|
+                  upper_limit = (upper_limit + len_count[i]) << 1
+                  @decode_len[i] = upper_limit << (16 - i)
+                  @decode_pos[i] = @decode_pos[i - 1] + len_count[i - 1]
+                end
+
+                # Fill decode_num
+                decode_pos_copy = @decode_pos.dup
+                bit_lengths.each_with_index do |len, symbol|
+                  next unless len.positive? && len < 16
+
+                  pos = decode_pos_copy[len]
+                  @decode_num[pos] = symbol if pos < @size
+                  decode_pos_copy[len] += 1
+                end
+
+                # Build quick lookup table
+                @quick_bits = 10 # Use 10 bits for quick lookup
+                build_quick_table(bit_lengths)
+
+                true
+              end
+
+              # Decode a symbol from bit reader
+              def decode(bit_reader)
+                # Read 16 bits for lookup
+                bit_field = bit_reader.read_bits(16)
+                return nil if bit_field.nil?
+
+                # Quick lookup
+                if @quick_len[bit_field].positive?
+                  bit_reader.skip_bits(@quick_len[bit_field])
+                  return @quick_num[bit_field]
+                end
+
+                # Full decode
+                bits = 15
+                (1..14).each do |i|
+                  if bit_field < @decode_len[i]
+                    bits = i
+                    break
+                  end
+                end
+
+                bit_reader.skip_bits(bits)
+
+                dist = bit_field - @decode_len[bits - 1]
+                dist >>= (16 - bits)
+                pos = @decode_pos[bits] + dist
+
+                pos < @size ? @decode_num[pos] : 0
+              end
+
+              private
+
+              def build_quick_table(bit_lengths)
+                quick_bits = @quick_bits
+
+                # Find maximum bit length for quick table
+                bit_lengths.each_with_index do |len, symbol|
+                  next unless len.positive? && len <= quick_bits
+
+                  # Calculate code for this symbol
+                  code = 0
+                  (0...len).each do |_i|
+                    code = (code << 1) | 1 # Simplified - should use actual codes
+                  end
+
+                  # Fill quick table entries
+                  extra_bits = quick_bits - len
+                  (0...(1 << extra_bits)).each do |extra|
+                    index = (code << extra_bits) | extra
+                    next if index >= @quick_len.size
+
+                    @quick_len[index] = len
+                    @quick_num[index] = symbol
+                  end
+                end
+              end
+            end
+
+            # RAR5 LZSS Decoder
+            #
+            # Based on libarchive's do_uncompress_block()
+            #
+            class Decoder
+              def initialize(data, window_size)
+                @data = data.dup.force_encoding(Encoding::BINARY)
+                @window_size = window_size
+                @window_mask = window_size - 1
+                @window = "\x00" * window_size
+                @output = StringIO.new
+                @output.set_encoding(Encoding::BINARY)
+                @write_ptr = 0
+                @dist_cache = [0, 0, 0, 0] # Distance cache
+                @last_len = 0
+              end
+
+              # Decode the compressed data
+              #
+              # @param expected_size [Integer, nil] Expected uncompressed size
+              # @return [String] Decompressed data
+              def decode(expected_size = nil)
+                return "" if @data.empty?
+
+                @bit_reader = BitReader.new(@data)
+
+                # Parse block header
+                parse_block_header
+
+                return @output.string unless @table_present
+
+                # Parse Huffman tables
+                return @output.string unless parse_huffman_tables
+
+                # Decode data
+                decode_data(expected_size)
+
+                @output.string
+              end
+
+              private
+
+              def parse_block_header
+                flags = @bit_reader.read_bits(8)
+                return unless flags
+
+                @table_present = flags.anybits?(0x01)
+              end
+
+              def parse_huffman_tables
+                # Parse bit lengths for BC table (20 codes)
+                bit_lengths_bc = parse_bit_lengths(HUFF_BC)
+                return false unless bit_lengths_bc
+
+                # Build BC table
+                @table_bc = HuffmanTable.new(HUFF_BC)
+                @table_bc.build(bit_lengths_bc)
+
+                # Parse main table using BC table
+                table_data = Array.new(HUFF_TABLE_SIZE, 0)
+                idx = 0
+
+                while idx < HUFF_TABLE_SIZE
+                  num = @table_bc.decode(@bit_reader)
+                  return false if num.nil?
+
+                  if num < 16
+                    # Direct value
+                    table_data[idx] = num
+                    idx += 1
+                  elsif num < 18
+                    # Repeat previous code
+                    count = num == 16 ? @bit_reader.read_bits(3) + 3 : @bit_reader.read_bits(7) + 11
+                    return false if count.nil? || idx.zero?
+
+                    count.times do
+                      break if idx >= HUFF_TABLE_SIZE
+
+                      table_data[idx] = table_data[idx - 1]
+                      idx += 1
+                    end
+                  else
+                    # Fill with zeros
+                    count = num == 18 ? @bit_reader.read_bits(3) + 3 : @bit_reader.read_bits(7) + 11
+                    return false if count.nil?
+
+                    count.times do
+                      break if idx >= HUFF_TABLE_SIZE
+
+                      table_data[idx] = 0
+                      idx += 1
+                    end
+                  end
+                end
+
+                # Build individual tables
+                @table_ld = HuffmanTable.new(HUFF_NC)
+                @table_ld.build(table_data[0, HUFF_NC])
+
+                @table_dd = HuffmanTable.new(HUFF_DC)
+                @table_dd.build(table_data[HUFF_NC, HUFF_DC])
+
+                @table_ldd = HuffmanTable.new(HUFF_LDC)
+                @table_ldd.build(table_data[HUFF_NC + HUFF_DC, HUFF_LDC])
+
+                @table_rd = HuffmanTable.new(HUFF_RC)
+                @table_rd.build(table_data[HUFF_NC + HUFF_DC + HUFF_LDC,
+                                           HUFF_RC])
+
+                true
+              end
+
+              def parse_bit_lengths(count)
+                lengths = Array.new(count, 0)
+                idx = 0
+                0xF0
+                4
+
+                while idx < count
+                  byte = @bit_reader.read_bits(8)
+                  return nil if byte.nil?
+
+                  # This is a simplified version
+                  # The actual libarchive uses nibble-based RLE
+                  lengths[idx] = byte & 0x0F
+                  idx += 1
+                  break if idx >= count
+
+                  lengths[idx] = (byte >> 4) & 0x0F
+                  idx += 1
+                end
+
+                lengths
+              end
+
+              def decode_data(expected_size)
+                while !@bit_reader.end_of_data?(@data.bytesize) &&
+                    (expected_size.nil? || @output.pos < expected_size)
+
+                  num = @table_ld.decode(@bit_reader)
+                  break if num.nil?
+
+                  if num < 256
+                    # Literal byte
+                    write_byte(num)
+                  elsif num == 256
+                    # Filter - skip for now
+                    skip_filter
+                  elsif num == 257
+                    # Repeat last match
+                    if @last_len.positive?
+                      copy_string(@last_len, @dist_cache[0])
+                    end
+                  elsif num < 262
+                    # Use distance cache entry
+                    cache_idx = num - 258
+                    dist = dist_cache_touch(cache_idx)
+
+                    len_slot = @table_rd.decode(@bit_reader)
+                    break if len_slot.nil?
+
+                    len = decode_code_length(len_slot)
+                    copy_string(len, dist) if len.positive?
+                  else
+                    # Regular match
+                    len = decode_code_length(num - 262)
+                    break if len <= 0
+
+                    dist_slot = @table_dd.decode(@bit_reader)
+                    break if dist_slot.nil?
+
+                    dist = decode_distance(dist_slot)
+                    break if dist <= 0
+
+                    dist_cache_push(dist)
+                    @last_len = len
+                    copy_string(len, dist)
+                  end
+                end
+              end
+
+              def decode_code_length(slot)
+                return slot + MIN_MATCH if slot < 16
+
+                # Extended length encoding
+                extra_bits = (slot - 12) / 2
+                base = ((2 + (slot & 1)) << extra_bits)
+                extra = @bit_reader.read_bits(extra_bits)
+                return 0 if extra.nil?
+
+                base + extra + MIN_MATCH
+              end
+
+              def decode_distance(slot)
+                return 0 if slot >= 64
+
+                if slot < 4
+                  return slot + 1
+                end
+
+                dbits = (slot / 2) - 1
+                dist = 2 | (slot & 1)
+                dist = (dist << dbits) | (1 << dbits)
+
+                if dbits >= 4
+                  # Read extra bits and low distance
+                  add = @bit_reader.read_bits_32(dbits - 4)
+                  dist += add << 4 if add
+
+                  low_dist = @table_ldd.decode(@bit_reader)
+                  return 0 if low_dist.nil?
+
+                  dist += low_dist
+                elsif dbits.positive?
+                  add = @bit_reader.read_bits(dbits)
+                  dist += add if add
+                end
+
+                # Adjust length based on distance
+                dist
+              end
+
+              def dist_cache_push(dist)
+                @dist_cache[3] = @dist_cache[2]
+                @dist_cache[2] = @dist_cache[1]
+                @dist_cache[1] = @dist_cache[0]
+                @dist_cache[0] = dist
+              end
+
+              def dist_cache_touch(idx)
+                dist = @dist_cache[idx]
+                if idx.positive?
+                  # Move to front
+                  (idx...DIST_CACHE_SIZE).each do |i|
+                    if i + 1 < DIST_CACHE_SIZE
+                      @dist_cache[i] =
+                        @dist_cache[i + 1]
+                    end
+                  end
+                  @dist_cache[0] = dist
+                end
+                dist
+              end
+
+              def write_byte(byte)
+                @output.putc(byte)
+                @window[@write_ptr & @window_mask] = byte.chr
+                @write_ptr += 1
+              end
+
+              def copy_string(length, distance)
+                return if distance <= 0 || distance > @write_ptr
+
+                length.times do
+                  read_idx = (@write_ptr - distance) & @window_mask
+                  byte = @window.getbyte(read_idx)
+                  write_byte(byte)
+                end
+              end
+
+              def skip_filter
+                # Skip filter data - simplified implementation
+                @bit_reader.skip_bits(16)
+              end
+            end
+          end
+        end
+      end
+    end
+  end
+end

data/lib/omnizip/formats/rar/rar5/compression/store.rb

@@ -0,0 +1,60 @@
+# frozen_string_literal: true
+
+module Omnizip
+  module Formats
+    module Rar
+      module Rar5
+        module Compression
+          # STORE compression method (uncompressed)
+          #
+          # This is the simplest compression method - it stores data without
+          # any compression. The "compressed" size equals the original size.
+          #
+          # @example Compress data
+          #   compressed = Store.compress("Hello, World!")
+          #   compressed # => "Hello, World!"
+          class Store
+            # Compression method identifier
+            METHOD = 0
+
+            # Compress data (passthrough for STORE)
+            #
+            # @param data [String] Data to compress
+            # @param _options [Hash] Options (ignored for STORE)
+            # @return [String] Uncompressed data
+            def self.compress(data, _options = {})
+              data
+            end
+
+            # Decompress data (passthrough for STORE)
+            #
+            # @param data [String] Data to decompress
+            # @param _options [Hash] Options (ignored for STORE)
+            # @return [String] Original data
+            def self.decompress(data, _options = {})
+              data
+            end
+
+            # Get compression method identifier
+            #
+            # @return [Integer] Method ID (0 for STORE)
+            def self.method_id
+              METHOD
+            end
+
+            # Get compression info VINT value
+            #
+            # For STORE, this is just the method ID (0)
+            # Bits 0-5: method (0=STORE)
+            # Bits 6+: version (0 for STORE)
+            #
+            # @return [Integer] Compression info value
+            def self.compression_info
+              METHOD
+            end
+          end
+        end
+      end
+    end
+  end
+end

data/lib/omnizip/formats/rar/rar5/crc32.rb

@@ -0,0 +1,39 @@
+# frozen_string_literal: true
+
+module Omnizip
+  module Formats
+    module Rar
+      module Rar5
+        # CRC32 calculation for RAR5 format
+        class CRC32
+          POLYNOMIAL = 0xEDB88320
+
+          # Generate CRC32 lookup table
+          def self.generate_table
+            @generate_table ||= (0..255).map do |i|
+              crc = i
+              8.times do
+                crc = (crc >> 1) ^ ((crc & 1) * POLYNOMIAL)
+              end
+              crc
+            end
+          end
+
+          # Calculate CRC32 for data
+          # @param data [String] Binary data
+          # @return [Integer] 32-bit CRC
+          def self.calculate(data)
+            table = generate_table
+            crc = 0xFFFFFFFF
+
+            data.bytes.each do |byte|
+              crc = (crc >> 8) ^ table[(crc ^ byte) & 0xFF]
+            end
+
+            crc ^ 0xFFFFFFFF
+          end
+        end
+      end
+    end
+  end
+end