omnizip 0.3.20 → 0.3.21

data/lib/omnizip/algorithms/zstandard/match_finder.rb

@@ -178,15 +178,18 @@ module Omnizip
           limit = block_end - mm
 
           while ip < limit
-            match = if ldm
+            match = if ldm || lazy.positive?
+                      # Lazy levels share the LDM loop's rep0 fast-path
+                      # (with backward extension): without it, levels
+                      # 6+ measured worse than the greedy default level
+                      # because every rep-offset match paid full offset
+                      # coding.
                       find_match_ldm(src, ip, ms, mm, limit, anchor,
                                      seq_store.rep_offsets[0], ldm,
                                      max_distance)
-                    elsif lazy.zero?
+                    else
                       find_greedy_match(src, ip, ms, mm, limit, anchor,
                                         seq_store.rep_offsets[0])
-                    else
-                      find_best_match(src, ip, ms, mm, limit)
                     end
             if match
               dist, len = match
@@ -247,7 +250,8 @@ module Omnizip
                                       ldm, max_distance)
           return nil unless dist
 
-          [dist, len, ip]
+          back = backward_extension(src, ip, anchor, ip - dist)
+          [dist, len + back, ip - back]
         end
 
         # Greedy-path match search (port of the Rust

data/lib/omnizip/algorithms/zstandard.rb

@@ -116,6 +116,7 @@ module Omnizip
         level = case options
                 when nil then nil
                 when Hash then options[:level]
+                # allowed: options is a public parameter; any object with #level is honored
                 else options.level if options.respond_to?(:level)
                 end
 

data/lib/omnizip/chunked/writer.rb

@@ -1,5 +1,8 @@
 # frozen_string_literal: true
 
+require "fileutils"
+require "stringio"
+
 module Omnizip
   module Chunked
     # Write large files incrementally in chunks

data/lib/omnizip/commands/archive_extract_command.rb

@@ -16,6 +16,8 @@
 # See the COPYING file for the complete text of the license.
 #
 
+require "fileutils"
+
 module Omnizip
   module Commands
     # Command to extract .7z archives to a directory.
@@ -222,7 +224,7 @@ module Omnizip
 
         extracted.size
       ensure
-        archive&.close if archive.respond_to?(:close)
+        archive.close if archive.is_a?(Omnizip::Zip::File)
       end
 
       def extract_gzip(archive_file, output_dir, verbose)

data/lib/omnizip/commands/archive_list_command.rb

@@ -69,7 +69,7 @@ module Omnizip
         entries = if patterns || excludes
                     filter_entries(archive, patterns, excludes)
                   else
-                    archive.respond_to?(:entries) ? archive.entries : archive.to_a
+                    archive.entries
                   end
 
         if count_only
@@ -91,6 +91,8 @@ module Omnizip
       rescue StandardError => e
         raise Omnizip::CompressionError,
               "Failed to list archive: #{e.message}"
+      ensure
+        archive.close if archive.is_a?(Omnizip::Zip::File)
       end
 
       def filter_entries(archive, patterns, excludes)
@@ -102,8 +104,7 @@ module Omnizip
         # Add exclude patterns
         excludes&.each { |pattern| filter.exclude_pattern(pattern) }
 
-        entries = archive.respond_to?(:entries) ? archive.entries : archive.to_a
-        filter.filter(entries)
+        filter.filter(archive.entries)
       end
 
       def display_simple_listing_filtered(entries)

data/lib/omnizip/commands/profile_show_command.rb

@@ -33,6 +33,8 @@ module Omnizip
         puts "Solid:       #{profile.solid}"
         puts "Description: #{profile.description}"
 
+        # ProfileRegistry accepts any CompressionProfile subclass.
+        # allowed: a registered subclass may expose base_profile
         return unless profile.respond_to?(:base_profile) && profile.base_profile
 
         puts "Based on:    #{profile.base_profile.name}"

data/lib/omnizip/convenience.rb

@@ -123,6 +123,7 @@ module Omnizip
       end
 
       handler = Omnizip::ArchiveHandler.for(format)
+      # allowed: handler is a registered duck; missing add_entry raises below
       if handler.respond_to?(:add_entry)
         handler.add_entry(archive_path, entry_name, source_path)
       else
@@ -142,6 +143,7 @@ module Omnizip
     def remove_from_archive(archive_path, entry_name, format: DEFAULT_FORMAT)
       require_archive!(archive_path)
       handler = Omnizip::ArchiveHandler.for(format)
+      # allowed: handler is a registered duck; missing remove_entry raises below
       unless handler.respond_to?(:remove_entry)
         raise Omnizip::UnsupportedFormatError,
               "Format #{format.inspect} does not support removing entries"

data/lib/omnizip/extraction/filter_chain.rb

@@ -159,10 +159,14 @@ module Omnizip
           entry
         else
           # Try common filename methods
+          # Narrowing this changes which files a filter matches.
+          # allowed: entries may be foreign objects
           if entry.respond_to?(:name)
             entry.name
+          # allowed: entries may be foreign objects
           elsif entry.respond_to?(:path)
             entry.path
+          # allowed: entries may be foreign objects
           elsif entry.respond_to?(:filename)
             entry.filename
           else

data/lib/omnizip/extraction/selective_extractor.rb

@@ -1,5 +1,7 @@
 # frozen_string_literal: true
 
+require "fileutils"
+
 module Omnizip
   module Extraction
     # Coordinates selective extraction from archives
@@ -120,8 +122,10 @@ module Omnizip
       #
       # @return [Array] All entries
       def list_all
+        # allowed: public API takes any archive object, including foreign ones
         if @archive.respond_to?(:entries)
           @archive.entries
+        # allowed: public API takes any Enumerable archive
         elsif @archive.respond_to?(:each)
           @archive.to_a
         else
@@ -150,10 +154,13 @@ module Omnizip
       # @param entry [Object] Entry to read
       # @return [String] Entry content
       def read_entry_content(entry)
+        # allowed: entries come from a caller-supplied archive
         if entry.respond_to?(:read)
           entry.read
+        # allowed: rubyzip-style entries expose get_input_stream, not read
         elsif entry.respond_to?(:get_input_stream)
           entry.get_input_stream.read
+        # allowed: some archives read entry content from the archive
         elsif @archive.respond_to?(:read)
           @archive.read(entry)
         else

data/lib/omnizip/file_type.rb

@@ -1,6 +1,7 @@
 # frozen_string_literal: true
 
 require "marcel"
+require "stringio"
 
 module Omnizip
   # File type detection module using Marcel for MIME type detection
@@ -95,16 +96,19 @@ module Omnizip
         return nil unless io
 
         # Save current position
+        # allowed: caller-supplied IO; only saved when the stream tracks one
         original_pos = io.pos if io.respond_to?(:pos)
 
         mime_type = Marcel::MimeType.for(io, name: filename)
 
         # Restore position
+        # allowed: caller-supplied IO capability probe
         io.seek(original_pos) if original_pos && io.respond_to?(:seek)
 
         mime_type
       rescue StandardError
         # Attempt to restore position even on error
+        # allowed: caller-supplied IO capability probe
         io.seek(original_pos) if original_pos && io.respond_to?(:seek)
         nil
       end

data/lib/omnizip/formats/ole.rb

@@ -1,5 +1,7 @@
 # frozen_string_literal: true
 
+require "fileutils"
+
 module Omnizip
   module Formats
     # OLE compound document format support

data/lib/omnizip/formats/rar/decompressor.rb

@@ -1,6 +1,7 @@
 # frozen_string_literal: true
 
 require "English"
+require "fileutils"
 require "tmpdir"
 module Omnizip
   module Formats

data/lib/omnizip/formats/rar/license_validator.rb

@@ -1,5 +1,7 @@
 # frozen_string_literal: true
 
+require "fileutils"
+
 module Omnizip
   module Formats
     module Rar

data/lib/omnizip/formats/rar/reader.rb

@@ -1,6 +1,7 @@
 # frozen_string_literal: true
 
 require "fileutils"
+require "stringio"
 
 module Omnizip
   module Formats
@@ -287,8 +288,6 @@ module Omnizip
         # @param entry [Models::RarEntry] File entry
         # @return [StringIO] Compressed data stream
         def read_compressed_data(entry)
-          require "stringio"
-
           # Find the entry's data offset in the archive
           File.open(@file_path, "rb") do |io|
             # Skip signature and headers

data/lib/omnizip/formats/rpm.rb

@@ -1,5 +1,6 @@
 # frozen_string_literal: true
 
+require "fileutils"
 require "stringio"
 require "tempfile"
 

data/lib/omnizip/formats/seven_zip/writer.rb

@@ -490,7 +490,15 @@ num_files)
 
         def compress_with_lzma2(data)
           # Use 7-Zip SDK LZMA2 encoder for 7-Zip format
+          #
+          # The encoder dictionary is sized to the data but never
+          # exceeds the size announced in the coder properties
+          # (options default 8MB): the decoder allocates its window
+          # from the announced size, so a larger encoder dictionary
+          # could emit matches reading outside the decoder window.
+          announced = @options[:dict_size] || (8 * 1024 * 1024)
           dict_size = [4096, data.bytesize].max
+          dict_size = announced if announced < dict_size
 
           encoder = Omnizip::Implementations::SevenZip::LZMA2::Encoder.new(
             dict_size: dict_size,

data/lib/omnizip/formats/tar/reader.rb

@@ -1,5 +1,7 @@
 # frozen_string_literal: true
 
+require "fileutils"
+
 module Omnizip
   module Formats
     module Tar

data/lib/omnizip/formats/xar/reader.rb

@@ -2,6 +2,8 @@
 
 require "zlib"
 require "digest"
+require "fileutils"
+require "stringio"
 
 module Omnizip
   module Formats

data/lib/omnizip/formats/xar/writer.rb

@@ -3,6 +3,7 @@
 require "zlib"
 require "digest"
 require "fileutils"
+require "stringio"
 
 module Omnizip
   module Formats

data/lib/omnizip/formats/xz.rb

@@ -1,5 +1,7 @@
 # frozen_string_literal: true
 
+require "stringio"
+
 module Omnizip
   module Formats
     # XZ compression format

data/lib/omnizip/formats/xz_impl/block_decoder.rb

@@ -77,6 +77,7 @@ module Omnizip
           end
 
           def set_encoding(enc)
+            # allowed: caller-supplied IO; String streams lack set_encoding
             @stream.set_encoding(enc) if @stream.respond_to?(:set_encoding)
           end
         end
@@ -110,7 +111,9 @@ module Omnizip
 
           if ENV["XZ_BLOCK_DEBUG"]
             warn "DEBUG: decode - compressed_size=#{compressed_size.inspect}, check_type=#{@check_type}"
+            # allowed: debug output under XZ_BLOCK_DEBUG
             warn "DEBUG: @input.class=#{@input.class}, @input.respond_to?(:pos)=#{@input.respond_to?(:pos)}"
+            # allowed: debug output under XZ_BLOCK_DEBUG
             pos = @input.respond_to?(:pos) ? @input.pos : "N/A"
             warn "DEBUG: @input.pos=#{pos}"
           end

data/lib/omnizip/formats/xz_impl/stream_decoder.rb

@@ -46,6 +46,7 @@ module Omnizip
 
           # Store original input and file size for backward_size validation (if available)
           original_input = input
+          # allowed: caller-supplied IO; only sized streams get validated
           original_file_size = input.size if input.respond_to?(:size)
 
           output, block_count, final_input, block_sizes = decode_blocks(input,
@@ -57,6 +58,7 @@ module Omnizip
           # multiples of four bytes...If the stored value does not match the real size of
           # the Index field, the decoder MUST indicate an error."
           # Reference: /Users/mulgogi/src/external/xz/src/liblzma/common/stream_decoder.c
+          # allowed: caller-supplied IO; footer check needs seeking
           if original_input.respond_to?(:seek) && original_file_size&.positive?
             # Use original input and file size for validation
             validate_backward_size_from_footer(original_input,
@@ -138,6 +140,7 @@ module Omnizip
         # @param byte [Integer] Byte to restore
         # @raise [RuntimeError] If IO doesn't support ungetbyte
         def self.restore_byte(input, byte)
+          # allowed: caller-supplied IO; FormatError follows when unsupported
           return input.ungetbyte(byte) if input.respond_to?(:ungetbyte)
 
           raise FormatError,
@@ -227,6 +230,7 @@ module Omnizip
         # @param check_type [Symbol] Expected checksum type
         # @param index_size [Integer, nil] Actual index size for backward_size validation
         def self.verify_footer_if_seekable(input, check_type, index_size = nil)
+          # allowed: caller-supplied IO; skipped on non-seekable streams
           return unless input.respond_to?(:seek) && input.respond_to?(:size) && input.size
 
           original_pos = input.pos
@@ -290,6 +294,7 @@ module Omnizip
         # @param input [IO] Input stream
         # @raise [FormatError] If there's invalid trailing data
         def self.verify_no_trailing_data(input)
+          # allowed: caller-supplied IO; needs byte-level positioning
           return unless input.respond_to?(:pos) && input.respond_to?(:getbyte)
 
           # Skip stream padding (null bytes)
@@ -304,6 +309,7 @@ module Omnizip
             else
               # Non-zero byte found - this should be a new stream or it's invalid
               # Restore the byte and check if it's a valid stream header
+              # allowed: caller-supplied IO capability probe
               input.ungetbyte(byte) if input.respond_to?(:ungetbyte)
 
               # Stream padding must be a multiple of 4 bytes
@@ -345,6 +351,7 @@ module Omnizip
           end
 
           # Restore the bytes we read
+          # allowed: caller-supplied IO capability probe
           if input.respond_to?(:ungetbyte)
             potential_header.reverse_each do |b|
               input.ungetbyte(b)
@@ -381,6 +388,7 @@ module Omnizip
         # @raise [FormatError] If backward_size points to invalid position
         def self.validate_backward_size_from_footer(input, file_size,
 _index_size)
+          # allowed: caller-supplied IO; skipped on non-seekable streams
           return unless input.respond_to?(:seek)
           return if file_size.nil? || file_size.zero?
 

data/lib/omnizip/formats/zip/reader.rb

@@ -1,5 +1,7 @@
 # frozen_string_literal: true
 
+require "fileutils"
+
 module Omnizip
   module Formats
     module Zip

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

@@ -121,6 +121,7 @@ module Omnizip
             # For raw mode, return actual decode bytes (excluding flush padding)
             if @raw_mode
               [@output.string, @range_encoder.bytes_for_decode]
+            # allowed: caller-supplied output; any #string sink buffers
             elsif @output.respond_to?(:string)
               # For File output, just return bytes written (don't try to read back)
               # For StringIO, return the string and its size

data/lib/omnizip/implementations/xz_utils/lzma2/decoder.rb

@@ -416,16 +416,11 @@ module Omnizip
 
                 if @lzma2_debug
                   warn "DEBUG: decompress_lzma_chunk - After set_input, checking range_decoder..."
-                  # Check if the decoder has a range_decoder variable
-                  if @lzma_decoder.respond_to?(:range_decoder)
-                    range_decoder = @lzma_decoder.range_decoder
-                    if range_decoder
-                      warn "  range_decoder exists: code=0x#{range_decoder.code.to_s(16)}, range=0x#{range_decoder.range.to_s(16)}, init_bytes_remaining=#{range_decoder.init_bytes_remaining}"
-                    else
-                      warn "  range_decoder is nil"
-                    end
+                  range_decoder = @lzma_decoder.range_decoder
+                  if range_decoder
+                    warn "  range_decoder exists: code=0x#{range_decoder.code.to_s(16)}, range=0x#{range_decoder.range.to_s(16)}, init_bytes_remaining=#{range_decoder.init_bytes_remaining}"
                   else
-                    warn "  @range_decoder not defined yet"
+                    warn "  range_decoder is nil"
                   end
                 end
               else

data/lib/omnizip/implementations/xz_utils/lzma2/encoder.rb

@@ -33,19 +33,6 @@ module Omnizip
         #
         # Ported from XZ Utils liblzma/lzma2_encoder.c
         #
-        # Compatibility helper for Ruby 3.0-3.1 where String#byteslice doesn't exist
-        module StringCompat
-          if "".respond_to?(:byteslice)
-            def self.byteslice(string, start, length)
-              string.byteslice(start, length)
-            end
-          else
-            def self.byteslice(string, start, length)
-              string.bytes[start, length]&.pack("C*") || ""
-            end
-          end
-        end
-
         # Constants
         UINT32_MAX = 0xFFFFFFFF
         REPS = 4
@@ -364,11 +351,9 @@ module Omnizip
 
             # Write to output stream
             if out_pos.value.positive?
-              # Use StringCompat.byteslice for Ruby 3.0-3.1 compatibility
               # Ruby's [] operator has a bug with null bytes that can return extra bytes
               # See: https://bugs.ruby-lang.org/issues/15985
-              output.write(StringCompat.byteslice(temp_buffer, 0,
-                                                  out_pos.value))
+              output.write(temp_buffer.byteslice(0, out_pos.value))
             end
 
             # Return the number of bytes written

data/lib/omnizip/io/buffered_input.rb

@@ -76,6 +76,7 @@ module Omnizip
       #
       # @return [void]
       def close
+        # allowed: caller-supplied source may not be closeable
         @source.close if @source.respond_to?(:close)
       end
 

data/lib/omnizip/io/buffered_output.rb

@@ -84,6 +84,7 @@ module Omnizip
       # @return [void]
       def close
         flush
+        # allowed: caller-supplied destination may not be closeable
         @destination.close if @destination.respond_to?(:close)
       end
 

data/lib/omnizip/io/source.rb

@@ -31,6 +31,7 @@ module Omnizip
         when ::IO, ::StringIO, ::Tempfile then new(input)
         when String then StringSource.new(input)
         else
+          # allowed: boundary validation; duck-typing ends at this adapter
           unless input.respond_to?(:read)
             raise ArgumentError,
                   "Cannot adapt #{input.inspect} to Omnizip::IO::Source"
@@ -56,6 +57,7 @@ module Omnizip
       end
 
       def close
+        # allowed: wrapped object may be a read-only duck with no close
         @io.close if @io.respond_to?(:close)
       end
 
@@ -98,6 +100,7 @@ module Omnizip
         when ::IO, ::StringIO, ::Tempfile then new(output)
         when String then PathSink.new(output)
         else
+          # allowed: boundary validation; duck-typing ends at this adapter
           unless output.respond_to?(:write)
             raise ArgumentError,
                   "Cannot adapt #{output.inspect} to Omnizip::IO::Sink"
@@ -116,6 +119,7 @@ module Omnizip
       end
 
       def close
+        # allowed: wrapped object may be a write-only duck with no close
         @io.close if @io.respond_to?(:close)
       end
 

data/lib/omnizip/io/stream_manager.rb

@@ -71,13 +71,14 @@ module Omnizip
       #
       # @return [void]
       def close
-        @source.close if @owned && @source.respond_to?(:close)
+        @source.close if @owned
       end
 
       # Check if source is at end.
       #
       # @return [Boolean] True if at EOF
       def eof?
+        # allowed: write-only sources have no eof?
         @source.eof? if @source.respond_to?(:eof?)
       end
 
@@ -103,6 +104,7 @@ module Omnizip
       end
 
       def validate_io_interface(source)
+        # allowed: boundary validation of a caller-supplied IO-like object
         if source.respond_to?(:read) || source.respond_to?(:write)
           source
         else

data/lib/omnizip/link_handler.rb

@@ -1,5 +1,7 @@
 # frozen_string_literal: true
 
+require "fileutils"
+
 module Omnizip
   # Handles symbolic and hard link operations with platform detection
   module LinkHandler

data/lib/omnizip/metadata/metadata_validator.rb

@@ -9,6 +9,8 @@ module Omnizip
       # @raise [ArgumentError] If metadata is invalid
       def validate_entry(metadata)
         validate_comment(metadata.comment) if metadata.comment
+        # validate_entry is public API.
+        # allowed: metadata need only expose what it validates
         validate_time(metadata.mtime) if metadata.respond_to?(:mtime)
         validate_permissions(metadata.unix_permissions) if metadata.unix_permissions
 

data/lib/omnizip/models/filter_config.rb

@@ -94,6 +94,8 @@ module Omnizip
       # @raise [NotImplementedError] If filter doesn't support id_for_format
       def id_for_format(format)
         filter = filter_instance
+        # FilterRegistry accepts any class, with no inheritance check.
+        # allowed: keeps NotImplementedError for filters without the method
         if filter.respond_to?(:id_for_format)
           filter.id_for_format(format)
         else

data/lib/omnizip/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Omnizip
-  VERSION = "0.3.20"
+  VERSION = "0.3.21"
 end

data/lib/omnizip/zip/file.rb

@@ -1,6 +1,7 @@
 # frozen_string_literal: true
 
 require "fileutils"
+require "stringio"
 
 module Omnizip
   module Zip
@@ -102,7 +103,6 @@ module Omnizip
         content = read_entry_data(entry)
 
         if block
-          require "stringio"
           StringIO.open(content, "rb", &block)
         else
           content

data/lib/omnizip/zip/input_stream.rb

@@ -1,5 +1,7 @@
 # frozen_string_literal: true
 
+require "stringio"
+
 module Omnizip
   module Zip
     # Rubyzip-compatible InputStream class
@@ -88,7 +90,6 @@ module Omnizip
             @current_entry.compression_method,
             @current_entry.size,
           )
-          require "stringio"
           @current_entry_io = StringIO.new(decompressed, "rb")
         end
 

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: omnizip
 version: !ruby/object:Gem::Version
-  version: 0.3.20
+  version: 0.3.21
 platform: ruby
 authors:
 - Ribose Inc.