omnizip 0.4.7-aarch64-linux

data/lib/omnizip/implementations/seven_zip/lzma/decoder.rb

@@ -0,0 +1,428 @@
+# frozen_string_literal: true
+
+# Copyright (C) 2025 Ribose Inc.
+#
+# Permission is hereby granted, free of charge, to any person obtaining a
+# copy of this software and associated documentation files (the "Software"),
+# to deal in the Software without restriction, including without limitation
+# the rights to use, copy, modify, merge, publish, distribute, sublicense,
+# and/or sell copies of the Software, and to permit persons to whom the
+# Software is furnished to do so, subject to the following conditions:
+#
+# The above copyright notice and this permission notice shall be included in
+# all copies or substantial portions of the Software.
+#
+# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+# AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+# FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+# DEALINGS IN THE SOFTWARE.
+
+module Omnizip
+  module Implementations
+    module SevenZip
+      module LZMA
+        # 7-Zip SDK compatible LZMA decoder.
+        #
+        # This decoder is designed to decode data encoded by the 7-Zip SDK encoder.
+        # Uses the same shared infrastructure (RangeDecoder, LengthCoder, DistanceCoder)
+        # to ensure model layout compatibility with the encoder.
+        class Decoder
+          include Omnizip::Algorithms::LZMA::Constants
+
+          # Maximum dictionary size to prevent memory exhaustion
+          # 64MB is a reasonable practical limit
+          MAX_DICT_SIZE = 64 * 1024 * 1024
+
+          attr_reader :lc, :lp, :pb, :dict_size, :uncompressed_size
+
+          # Initialize 7-Zip LZMA decoder
+          #
+          # @param input [IO] Input stream with LZMA compressed data
+          # @param options [Hash] Decoding options
+          # @option options [Boolean] :raw_mode Skip header parsing (for LZMA2)
+          # @option options [Integer] :lc Literal context bits
+          # @option options [Integer] :lp Literal position bits
+          # @option options [Integer] :pb Position bits
+          # @option options [Integer] :dict_size Dictionary size
+          def initialize(input, options = {})
+            @input = input
+            @raw_mode = options.fetch(:raw_mode, false)
+
+            if @raw_mode
+              @lc = options[:lc] || 3
+              @lp = options[:lp] || 0
+              @pb = options[:pb] || 2
+              @dict_size = [[options[:dict_size] || (1 << 16), 1].max,
+                            MAX_DICT_SIZE].min
+              @uncompressed_size = options[:uncompressed_size]
+            else
+              parse_header
+            end
+
+            init_decoder
+          end
+
+          # Decode LZMA stream
+          #
+          # @param output [IO, nil] Output stream (if nil, returns String)
+          # @param preserve_dict [Boolean] Preserve dictionary for chunked decoding
+          # @return [String, Integer] Decompressed data or bytes written
+          def decode_stream(output = nil, preserve_dict: false)
+            @output_buffer = []
+            unless preserve_dict && @dictionary
+              @dictionary = Array.new(@dict_size,
+                                      0)
+            end
+            @dict_pos = 0
+            @dict_full = false
+
+            # Initialize range decoder (7-Zip SDK version)
+            @range_decoder = RangeDecoder.new(@input)
+
+            # Main decode loop
+            loop do
+              break if reached_end?
+
+              # Track if we were using EOPM before decoding
+              was_using_eopm = @allow_eopm
+
+              decode_symbol
+
+              # If we were using EOPM and now we're not, EOS was detected
+              break if was_using_eopm && !@allow_eopm
+            end
+
+            result = @output_buffer.pack("C*").force_encoding(Encoding::BINARY)
+
+            if output
+              output.write(result)
+              result.bytesize
+            else
+              result
+            end
+          end
+
+          # Reset decoder state for new chunk (LZMA2)
+          def reset(new_lc: nil, new_lp: nil, new_pb: nil, preserve_dict: false)
+            @lc = new_lc if new_lc
+            @lp = new_lp if new_lp
+            @pb = new_pb if new_pb
+
+            unless preserve_dict
+              @dictionary = Array.new(@dict_size, 0)
+              @dict_pos = 0
+              @dict_full = false
+            end
+
+            init_decoder
+          end
+
+          # Set new input stream (LZMA2)
+          def set_input(new_input)
+            @input = new_input
+          end
+
+          # Set uncompressed size (LZMA2)
+          def set_uncompressed_size(size, allow_eopm: true)
+            @uncompressed_size = size
+            @allow_eopm = allow_eopm
+          end
+
+          private
+
+          # Parse LZMA header
+          def parse_header
+            props = @input.getbyte
+            raise Omnizip::DecompressionError, "Invalid LZMA header: missing properties" unless props
+
+            @lc = props % 9
+            rem = props / 9
+            @lp = rem % 5
+            @pb = rem / 5
+
+            raise Omnizip::DecompressionError, "Invalid LZMA properties: pb=#{@pb} > 4" if @pb > 4
+
+            @dict_size = 0
+            4.times do |i|
+              byte = @input.getbyte
+              raise Omnizip::DecompressionError, "Invalid LZMA header: missing dictionary size" unless byte
+
+              @dict_size |= (byte << (i * 8))
+            end
+
+            # Clamp dict_size to prevent memory exhaustion
+            @dict_size = [[@dict_size, 1].max, MAX_DICT_SIZE].min
+
+            @uncompressed_size = 0
+            8.times do |i|
+              byte = @input.getbyte
+              raise Omnizip::DecompressionError, "Invalid LZMA header: missing uncompressed size" unless byte
+
+              @uncompressed_size |= (byte * (1 << (i * 8)))
+            end
+
+            @allow_eopm = (@uncompressed_size == 0xFFFFFFFFFFFFFFFF)
+          end
+
+          # Initialize decoder state
+          def init_decoder
+            @state = StateMachine.new
+            @reps = [0, 0, 0, 0]
+
+            # Calculate literal_mask using XZ Utils formula
+            # literal_mask = (UINT32_C(0x100) << (lp)) - (UINT32_C(0x100) >> (lc))
+            @literal_mask = (0x100 << @lp) - (0x100 >> @lc)
+
+            # Initialize literal models using same layout as encoder
+            # max_context = literal_mask
+            # max_base_offset = 3 * (max_context << lc)
+            # max_model_index = max_base_offset + 0x300
+            max_context = @literal_mask
+            max_base_offset = 3 * (max_context << @lc)
+            max_model_index = max_base_offset + 0x300
+
+            @literal_models = Array.new(max_model_index + 1) do
+              Omnizip::Algorithms::LZMA::BitModel.new
+            end
+
+            # Initialize probability models using same layout as encoder
+            num_pos_states = 1 << @pb
+            @is_match_models = Array.new(NUM_STATES * num_pos_states) do
+              Omnizip::Algorithms::LZMA::BitModel.new
+            end
+            @is_rep_models = Array.new(NUM_STATES) do
+              Omnizip::Algorithms::LZMA::BitModel.new
+            end
+            @is_rep0_models = Array.new(NUM_STATES) do
+              Omnizip::Algorithms::LZMA::BitModel.new
+            end
+            @is_rep1_models = Array.new(NUM_STATES) do
+              Omnizip::Algorithms::LZMA::BitModel.new
+            end
+            @is_rep2_models = Array.new(NUM_STATES) do
+              Omnizip::Algorithms::LZMA::BitModel.new
+            end
+            @is_rep0_long_models = Array.new(NUM_STATES * num_pos_states) do
+              Omnizip::Algorithms::LZMA::BitModel.new
+            end
+
+            # Use shared LengthCoder and DistanceCoder (they have decode methods)
+            @length_coder = Omnizip::Algorithms::LZMA::LengthCoder.new(num_pos_states)
+            @rep_length_coder = Omnizip::Algorithms::LZMA::LengthCoder.new(num_pos_states)
+            @distance_coder = Omnizip::Algorithms::LZMA::DistanceCoder.new(NUM_LEN_TO_POS_STATES)
+
+            @output_count = 0
+          end
+
+          # Check if we've reached end of stream
+          def reached_end?
+            if @allow_eopm
+              false
+            else
+              @output_count >= @uncompressed_size
+            end
+          end
+
+          # Decode one symbol
+          def decode_symbol
+            pos_state = @output_count & ((1 << @pb) - 1)
+
+            # Decode is_match using same model layout as encoder
+            model_index = (@state.value * (1 << @pb)) + pos_state
+            is_match = @range_decoder.decode_bit(@is_match_models[model_index])
+
+            if is_match.zero?
+              decode_literal
+            else
+              decode_match(pos_state)
+            end
+          end
+
+          # Decode a literal byte
+          def decode_literal
+            prev_byte = @dict_pos.positive? ? @dictionary[@dict_pos - 1] : 0
+
+            # Calculate base_offset using XZ Utils formula (same as encoder)
+            # context = (((pos << 8) + prev_byte) & literal_mask)
+            # base_offset = 3 * (context << lc)
+            context = (((@output_count << 8) + prev_byte) & @literal_mask)
+            base_offset = 3 * (context << @lc)
+
+            symbol = 1
+
+            if @state.use_matched_literal?
+              # Matched literal: use match byte from dictionary
+              #
+              # XZ Utils decoder pattern (from range_decoder.h rc_matched_literal):
+              # - symbol starts at 1
+              # - match_byte is shifted FIRST, then match_bit extracted
+              # - subcoder_index = offset + match_bit + symbol (symbol not shifted)
+              # - offset updated: if bit 0: offset &= ~match_bit, if bit 1: offset &= match_bit
+              # - loop 8 times
+              #
+              match_byte_val = get_byte_from_dict(@reps[0])
+              offset = 0x100
+              symbol = 1 # XZ Utils starts at 1, not 0x100
+
+              8.times do
+                # Shift match_byte FIRST (matches XZ Utils decoder)
+                match_byte_val <<= 1
+
+                # Get match_bit from shifted value
+                match_bit = match_byte_val & offset
+
+                # Calculate subcoder index (XZ Utils: offset + match_bit + symbol)
+                model_idx = base_offset + offset + match_bit + symbol
+
+                # Decode bit
+                bit = @range_decoder.decode_bit(@literal_models[model_idx])
+
+                # Update symbol and offset based on decoded bit
+                if bit.zero?
+                  symbol <<= 1
+                  offset &= ~match_bit
+                else
+                  symbol = (symbol << 1) + 1
+                  offset &= match_bit
+                end
+              end
+
+            else
+              # Normal (unmatched) literal
+              8.times do
+                bit = @range_decoder.decode_bit(@literal_models[base_offset + symbol])
+                symbol = (symbol << 1) | bit
+              end
+            end
+            byte = symbol & 0xFF
+
+            put_byte_to_dict(byte)
+            @output_buffer << byte
+            @output_count += 1
+            @state.update_literal
+          end
+
+          # Decode a match
+          def decode_match(pos_state)
+            is_rep = @range_decoder.decode_bit(@is_rep_models[@state.value])
+
+            if is_rep.zero?
+              # Simple match
+              len = @length_coder.decode(@range_decoder,
+                                         pos_state) + MATCH_LEN_MIN
+              @state.update_match
+
+              # Decode distance
+              len_state = [len - MATCH_LEN_MIN, NUM_LEN_TO_POS_STATES - 1].min
+              distance = @distance_coder.decode(@range_decoder, len_state)
+
+              # Check for EOPM (distance = 0xFFFFFFFF means end marker)
+              if distance == 0xFFFFFFFF
+                @allow_eopm = false
+                return
+              end
+
+              # Decoder returns distance before +1, so add 1 to get actual distance
+              distance += 1
+
+              raise Omnizip::DecompressionError, "Invalid distance: #{distance}" if distance >= @dict_size && @dict_full
+
+              # Update reps
+              @reps[3] = @reps[2]
+              @reps[2] = @reps[1]
+              @reps[1] = @reps[0]
+              @reps[0] = distance
+            else
+              # Repeated match
+              len, distance = decode_rep_match(pos_state)
+            end
+
+            # Copy from dictionary
+            len.times do
+              byte = get_byte_from_dict(distance)
+              put_byte_to_dict(byte.ord)
+              @output_buffer << byte
+              @output_count += 1
+            end
+          end
+
+          # Decode repeated match
+          def decode_rep_match(pos_state)
+            if @range_decoder.decode_bit(@is_rep0_models[@state.value]).zero?
+              # Rep0
+              if @range_decoder.decode_bit(
+                @is_rep0_long_models[(@state.value * (1 << @pb)) + pos_state],
+              ).zero?
+                # Short rep (length 1)
+                @state.update_short_rep
+                return [1, @reps[0]]
+              end
+
+              len = @rep_length_coder.decode(@range_decoder,
+                                             pos_state) + MATCH_LEN_MIN
+              @state.update_rep
+              return [len, @reps[0]]
+            end
+
+            if @range_decoder.decode_bit(@is_rep1_models[@state.value]).zero?
+              # Rep1
+              len = @rep_length_coder.decode(@range_decoder,
+                                             pos_state) + MATCH_LEN_MIN
+              distance = @reps[1]
+              @reps[1] = @reps[0]
+              @reps[0] = distance
+              @state.update_rep
+              return [len, distance]
+            end
+
+            if @range_decoder.decode_bit(@is_rep2_models[@state.value]).zero?
+              # Rep2
+              len = @rep_length_coder.decode(@range_decoder,
+                                             pos_state) + MATCH_LEN_MIN
+              distance = @reps[2]
+              @reps[2] = @reps[1]
+              @reps[1] = @reps[0]
+              @reps[0] = distance
+              @state.update_rep
+              return [len, distance]
+            end
+
+            # Rep3
+            len = @rep_length_coder.decode(@range_decoder,
+                                           pos_state) + MATCH_LEN_MIN
+            distance = @reps[3]
+            @reps[3] = @reps[2]
+            @reps[2] = @reps[1]
+            @reps[1] = @reps[0]
+            @reps[0] = distance
+            @state.update_rep
+            [len, distance]
+          end
+
+          # Get byte from dictionary at distance
+          # Distance 1 = most recent byte = position dict_pos - 1
+          # Distance N = position dict_pos - N
+          def get_byte_from_dict(distance)
+            if distance > @dict_pos && !@dict_full
+              0
+            else
+              # Use positive modulo to handle negative numbers correctly
+              pos = (@dict_pos - distance) % @dict_size
+              @dictionary[pos]
+            end
+          end
+
+          # Put byte to dictionary
+          def put_byte_to_dict(byte)
+            @dictionary[@dict_pos] = byte
+            @dict_pos = (@dict_pos + 1) % @dict_size
+            @dict_full = true if @dict_pos.zero?
+          end
+        end
+      end
+    end
+  end
+end