omnizip 0.3.51 → 0.3.53

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 4b268c5b315ddcfff3347f6843e30e7ca180b1b3bae14b21a7a721bd2ccf3f6e
-  data.tar.gz: 2d150420264453af6e1ba5114b32a31b1a67c12a020f41435a037747dec2bfde
+  metadata.gz: a7e72a57eaeb6e597e40020de67aa0242da65c6e6e8ceaa933d779d926e55666
+  data.tar.gz: 105cf4f238610f518462c115d398403fef5daca9b2994788361b027c7c5ec648
 SHA512:
-  metadata.gz: '053812e04bf42b6fe5335d57a5183c90b19a3ed8b414ee57ac9a83831a04f15e7225ecc79d29bc42e78e692bf7fe6f60b88ba234a3fc92eb0e781e7a357acd14'
-  data.tar.gz: dd2e32bd6ea8732be35a2f6c7325ed23b922860c051caf7666df93edd7f7da1342bdc1f6c9928346a404a7a1c707c509bd74d1b281ddd07cdc4a591c67e8a05d
+  metadata.gz: a5301b14047d1150885624527d03176c61a979464b7a16329a47449e13fe05679c1e5ce11a650bfd29101aaaf502b5ab81513558fc080dca85554266ae30f8dc
+  data.tar.gz: a44ffd04cc02fa03e2b23a330da4f6612d268bfd350d9b9672e42d7a107dcad65201ed9707651b3533fdd790a5841a5ff9297d6f3a8e937b51e0f1614d756c55

data/CHANGELOG.md

@@ -7,6 +7,62 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+## [0.3.52] - 2026-09-01
+
+### Changed
+- The extract command's pattern path opens archives through the
+  Archive facade instead of a hand-rolled extension `case` that
+  misrouted `.cpio`/`.iso`/`.rpm` files into the 7z reader — pattern
+  extraction now works for every routed format. The metadata
+  command's 7z view lists through the handler registry (the detail
+  contract gains `:crc`), and SafeExtract extracts through
+  `extract_archive`, so every routed format gets rollback and
+  verification instead of ZIP-only with a `NotImplementedError`
+  for 7z. SelectiveExtractor understands both entry contracts
+  (legacy `entry_name` and reader-session `name`) and skips
+  directory entries when extracting.
+- `CliOutputFormatter.format_size` is now the single byte formatter;
+  the nine private `format_bytes`/`format_size` copies (four
+  different shapes, two unit ceilings) become delegators.
+
+### Fixed
+- The long-deferred parity spec flake: `Par2Verifier` crashed five
+  frames deep on a nil PAR2 path (`File.exist?(nil)` → TypeError)
+  where the repairer already raised a clean ArgumentError; the
+  verifier now mirrors it, and the spec's `.verify` group gains the
+  nil-index fallback `.repair` already had. The dead shadowed
+  `analyze_file` definition in `Par2Creator` (silently overridden
+  by a second definition with a divergent file_id format) is
+  deleted. The tautological "LZMA2 is implicit" xz assertion became
+  a real block-header parse asserting the explicit 0x21 filter ID.
+
+## [0.3.51] - 2026-09-01
+
+### Changed
+- Profile resolution lives in one place: `Omnizip::Profile` gains
+  `resolve`, `apply`, and `first_file_among`, and the convenience
+  methods plus the `archive create` command collapse onto them. The
+  two prior copies had drifted — the command only accepted `"auto"`
+  strings and could not pass a `CompressionProfile` through.
+- Both direct converter strategies (ZIP↔7z) now extract through
+  `Omnizip.extract_archive` and write through a shared
+  `ConversionStrategy#repack_tree` on the Archive facade — they were
+  the last library code instantiating format readers/writers
+  directly, and the 7z→ZIP path required `#open` against the
+  reader-always-usable invariant.
+- `create_rar` routes through the RAR handler (unrar-verified) with
+  the generic writer block interface; its stale "requires RAR
+  license" docstring is gone. `decompress_file`'s three-arm case
+  with magic `:xz`/`:zstandard` markers becomes a
+  `SINGLE_FILE_DECOMPRESSORS` table mirroring the compressor table.
+
+### Fixed
+- macOS CI failed at setup from 2026-09-01: Homebrew disabled the
+  `rar` cask (Gatekeeper check), breaking `brew install rar`. The
+  workflow now installs it separately and non-fatally, like the
+  Windows branch already did for winrar; RAR interop specs skip
+  when unrar is absent.
+
 ## [0.3.50] - 2026-08-31
 
 ### Added

data/lib/omnizip/archive_handlers/seven_zip_handler.rb

@@ -31,7 +31,7 @@ module Omnizip
             entries.map do |e|
               { name: e.name, size: e.size, directory: e.is_dir,
                 compressed_size: (e.compressed_size if e.compressed_size&.positive?),
-                mtime: e.mtime }
+                mtime: e.mtime, crc: e.crc }
             end
           else
             entries.map(&:name)

data/lib/omnizip/archive_handlers/zip_handler.rb

@@ -38,7 +38,7 @@ module Omnizip
           extracted = []
           reader.entries.each do |entry|
             dest_path = ::File.join(output_dir, entry.filename)
-            raise "File exists: #{dest_path}" if ::File.exist?(dest_path)
+            raise Errno::EEXIST, "File exists: #{dest_path}" if ::File.exist?(dest_path)
 
             reader.extract_entry(entry, output_dir)
             extracted << dest_path

data/lib/omnizip/buffer/memory_extractor.rb

@@ -151,24 +151,12 @@ module Omnizip
 
       private
 
-      # Detect archive format from magic bytes
+      # Delegate to the shared buffer-side sniffer
       #
       # @return [Symbol] Detected format
       # @raise [Omnizip::FormatError] If format cannot be detected
       def detect_format
-        magic = @buffer.read(4).to_s.b
-        @buffer.rewind
-
-        case magic
-        when "PK\x03\x04".b, "PK\x05\x06".b, "PK\x07\x08".b
-          # ZIP signatures: local file header, EOCD, data descriptor
-          :zip
-        when "7z\xBC\xAF".b
-          :seven_zip
-        else
-          raise Omnizip::FormatError,
-                "Unknown archive format (magic: #{magic.inspect})"
-        end
+        Omnizip::Buffer.detect_format(@buffer)
       end
 
       # Extract all entries from ZIP
@@ -207,8 +195,7 @@ module Omnizip
       def extract_entry_seven_zip(name)
         @buffer.rewind
         SevenZipBridge.open(@buffer) do |archive|
-          entry = archive.raw_entries.find { |e| e.name == name }
-          archive.extract_all_to_memory[name] if entry
+          archive.read_entry(name)
         end
       end
 

data/lib/omnizip/buffer/seven_zip_bridge.rb

@@ -112,6 +112,19 @@ module Omnizip
           @reader.list_files
         end
 
+        # Content of a single named file entry, or nil when absent
+        # (directories carry no content). Extracts just that entry —
+        # not the whole archive.
+        #
+        # @param name [String] Entry name
+        # @return [String, nil]
+        def read_entry(name)
+          entry = raw_entries.find { |e| e.name == name && !e.is_dir }
+          return nil unless entry
+
+          Entry.new(entry, @reader, @tmp).read
+        end
+
         # Entry wrapper matching MemoryArchive::Entry's surface
         class Entry
           attr_reader :name, :size

data/lib/omnizip/buffer.rb

@@ -107,27 +107,9 @@ module Omnizip
         extractor.extract_all
       end
 
-      # Create archive from Hash of filename => content
-      #
-      # @param hash [Hash<String, String>] Filename => content mapping
-      # @param format [Symbol] Archive format
-      # @param options [Hash] Format-specific options
-      # @return [StringIO] Complete archive in memory
-      #
-      # @example Create from Hash
-      #   data = {'file1.txt' => 'content1', 'file2.txt' => 'content2'}
-      #   zip = Omnizip::Buffer.create_from_hash(data, :zip)
-      def create_from_hash(hash, format = :zip, **options)
-        create(format, **options) do |archive|
-          hash.each do |name, content|
-            archive.add(name, content)
-          end
-        end
-      end
-
-      private
-
-      # Detect archive format from magic bytes
+      # Detect archive format from magic bytes. THE buffer-side
+      # sniffer — in-memory data has no extension to route by, so
+      # every consumer delegates here.
       #
       # @param buffer [StringIO] Buffer containing archive data
       # @return [Symbol] Detected format
@@ -148,6 +130,26 @@ module Omnizip
         end
       end
 
+      # Create archive from Hash of filename => content
+      #
+      # @param hash [Hash<String, String>] Filename => content mapping
+      # @param format [Symbol] Archive format
+      # @param options [Hash] Format-specific options
+      # @return [StringIO] Complete archive in memory
+      #
+      # @example Create from Hash
+      #   data = {'file1.txt' => 'content1', 'file2.txt' => 'content2'}
+      #   zip = Omnizip::Buffer.create_from_hash(data, :zip)
+      def create_from_hash(hash, format = :zip, **options)
+        create(format, **options) do |archive|
+          hash.each do |name, content|
+            archive.add(name, content)
+          end
+        end
+      end
+
+      private
+
       # Create ZIP archive in buffer
       #
       # @param buffer [StringIO] Buffer to write to

data/lib/omnizip/checksums/verifier.rb

@@ -94,7 +94,7 @@ module Omnizip
         when CHECK_SHA256
           verify_sha256(data, expected)
         else
-          raise "Unknown check type: #{check_type}"
+          raise Omnizip::UnknownChecksumError, "Unknown check type: #{check_type}"
         end
       end
 
@@ -122,7 +122,7 @@ module Omnizip
         when CHECK_SHA256
           Digest::SHA256.digest(data)
         else
-          raise "Unknown check type: #{check_type}"
+          raise Omnizip::UnknownChecksumError, "Unknown check type: #{check_type}"
         end
       end
     end

data/lib/omnizip/chunked.rb

@@ -95,30 +95,23 @@ module Omnizip
                 "Input archive not found: #{input}"
         end
 
-        writer = Writer.new(output, chunk_size: chunk_size)
-        processed = 0
-
-        Omnizip::Zip::File.open(input) do |zip|
-          entry = zip.entries.first
-          total_size = entry.size
-
-          # Read the full entry content
-          content = zip.get_input_stream(entry)
+        reader = Omnizip::Formats::Zip::Reader.new(input)
+        entry = reader.entries.first
+        raise Omnizip::Error, "No entries to decompress in #{input}" unless entry
 
-          # Write in chunks
-          offset = 0
-          while offset < content.bytesize
-            chunk = content.byteslice(offset, chunk_size) || ""
-            break if chunk.empty?
+        total_size = entry.uncompressed_size.to_i
+        processed = 0
+        writer = Writer.new(output, chunk_size: chunk_size)
 
-            writer.write_chunk(chunk)
-            processed += chunk.bytesize
-            offset += chunk.bytesize
+        # Streams entry-by-entry in bounded memory — the whole entry
+        # is never held at once
+        reader.read_entry_stream(entry.filename, chunk_size: chunk_size) do |chunk|
+          writer.write_chunk(chunk)
+          processed += chunk.bytesize
 
-            if progress
-              percentage = (processed.to_f / total_size * 100).round(2)
-              progress.call(processed, total_size, percentage)
-            end
+          if progress
+            percentage = (processed.to_f / total_size * 100).round(2)
+            progress.call(processed, total_size, percentage)
           end
         end
 

data/lib/omnizip/cli/output_formatter.rb

@@ -44,19 +44,27 @@ module Omnizip
         ].join("\n")
       end
 
-      # Format file size in human-readable format.
+      # Format file size in human-readable format. This is THE byte
+      # formatter — commands, the CLI, and display-oriented models
+      # delegate here instead of carrying private copies.
       #
-      # @param bytes [Integer] Size in bytes
-      # @return [String] Formatted size
+      # @param bytes [Integer, nil] Size in bytes
+      # @return [String] Formatted size ("512 B", "1.5 KB", "2.0 MB")
       def format_size(bytes)
+        return "0 B" if bytes.nil? || bytes.zero?
+
         units = %w[B KB MB GB TB]
-        return "0 B" if bytes.zero?
+        size = bytes.to_f
+        unit_index = 0
+
+        while size >= 1024.0 && unit_index < units.size - 1
+          size /= 1024.0
+          unit_index += 1
+        end
 
-        exp = (Math.log(bytes) / Math.log(1024)).floor
-        exp = [exp, units.length - 1].min
+        return format("%d B", bytes) if unit_index.zero?
 
-        size = (bytes.to_f / (1024**exp)).round(2)
-        "#{size} #{units[exp]}"
+        format("%.1f %s", size, units[unit_index])
       end
 
       # Format error message for display.

data/lib/omnizip/cli.rb

@@ -60,13 +60,7 @@ module Omnizip
     # @param bytes [Integer] the byte count
     # @return [String] human-readable size (e.g. "1.5 MB")
     def format_bytes(bytes)
-      return "0 B" if bytes.zero?
-
-      units = %w[B KB MB GB TB]
-      exp = (Math.log(bytes) / Math.log(1024)).to_i
-      exp = [exp, units.size - 1].min
-
-      "%.1f %s" % [bytes.to_f / (1024**exp), units[exp]]
+      Omnizip::CliOutputFormatter.format_size(bytes)
     end
   end
 

data/lib/omnizip/commands/archive_create_command.rb

@@ -201,16 +201,7 @@ module Omnizip
       end
 
       def format_bytes(bytes)
-        units = %w[B KB MB GB]
-        size = bytes.to_f
-        unit_idx = 0
-
-        while size >= 1024 && unit_idx < units.length - 1
-          size /= 1024.0
-          unit_idx += 1
-        end
-
-        format("%.2f %s", size, units[unit_idx])
+        Omnizip::CliOutputFormatter.format_size(bytes)
       end
 
       def parse_volume_size(size_str)

data/lib/omnizip/commands/archive_extract_command.rb

@@ -135,17 +135,10 @@ module Omnizip
       end
 
       def extract_with_patterns(archive_file, output_dir, verbose)
-        # Determine archive type and open appropriately
-        archive = case File.extname(archive_file).downcase
-                  when ".zip"
-                    Omnizip::Zip::File.open(archive_file)
-                  when ".rar"
-                    Formats::Rar::Reader.new(archive_file).open
-                  when ".tar"
-                    Formats::Tar::Reader.new(archive_file).read
-                  else
-                    Formats::SevenZip::Reader.new(archive_file).open
-                  end
+        # Route through the facade: the extension routes pick each
+        # format's reader, and the reader session holds no open
+        # resources to close
+        archive = Omnizip::Archive.open(archive_file)
 
         # Build filter chain
         filter = Extraction::FilterChain.new
@@ -193,8 +186,6 @@ module Omnizip
         end
 
         extracted.size
-      ensure
-        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

@@ -157,24 +157,7 @@ module Omnizip
       end
 
       def format_bytes(bytes)
-        return "0 B" if bytes.zero?
-
-        units = %w[B KB MB GB]
-        size = bytes.to_f
-        unit_idx = 0
-
-        while size >= 1024 && unit_idx < units.length - 1
-          size /= 1024.0
-          unit_idx += 1
-        end
-
-        if size < 10
-          format("%.2f %s", size, units[unit_idx])
-        elsif size < 100
-          format("%.1f %s", size, units[unit_idx])
-        else
-          format("%.0f %s", size, units[unit_idx])
-        end
+        Omnizip::CliOutputFormatter.format_size(bytes)
       end
 
       # Helper methods to handle different entry types. Each delegates

data/lib/omnizip/commands/metadata_command.rb

@@ -54,7 +54,8 @@ module Omnizip
         %i[comment set_mtime chmod set_attribute].none? { |key| options[key] }
       end
 
-      # Read-only view for 7z archives; the edit path is zip-only
+      # Read-only view for 7z archives through the handler registry;
+      # the edit path is zip-only
       def show_seven_zip_metadata(archive_path, pattern)
         if options[:comment] || options[:set_mtime] ||
             options[:chmod] || options[:set_attribute]
@@ -62,37 +63,37 @@ module Omnizip
                 "Metadata editing is only supported for ZIP archives"
         end
 
-        reader = Omnizip::Formats::SevenZip::Reader.new(archive_path)
-        reader.open
+        entries = Omnizip.list_archive(archive_path, details: true)
 
         if pattern
-          entries = reader.list_files.select do |e|
-            File.fnmatch(pattern, e.name)
+          entries = entries.select do |entry|
+            File.fnmatch(pattern, entry[:name])
           end
           if entries.empty?
             warn "No entries match pattern: #{pattern}"
             return
           end
 
-          entries.each do |entry|
-            puts "Entry: #{entry.name}"
-            puts "  Type: #{entry.is_dir ? 'Directory' : 'File'}"
-            puts "  Size: #{format_size(entry.size)}"
-            # Per-entry packed sizes are not tracked by the parser
-            if entry.compressed_size&.positive?
-              puts "  Compressed: #{format_size(entry.compressed_size)}"
-            end
-            puts "  Modified: #{entry.mtime}" if entry.mtime
-            puts "  CRC32: 0x#{entry.crc.to_s(16).upcase}" if entry.crc
-          end
+          entries.each { |entry| show_seven_zip_entry(entry) }
         else
-          entries = reader.list_files
-          files = entries.reject(&:is_dir)
+          files = entries.reject { |entry| entry[:directory] }
           puts "Archive: #{archive_path}"
           puts "Entries: #{entries.size} " \
                "(#{files.size} files, #{entries.size - files.size} dirs)"
-          puts "Total size: #{format_size(files.sum(&:size))}"
+          puts "Total size: #{format_size(files.sum { |entry| entry[:size] })}"
+        end
+      end
+
+      def show_seven_zip_entry(entry)
+        puts "Entry: #{entry[:name]}"
+        puts "  Type: #{entry[:directory] ? 'Directory' : 'File'}"
+        puts "  Size: #{format_size(entry[:size])}"
+        # Per-entry packed sizes are not tracked by the parser
+        if entry[:compressed_size]&.positive?
+          puts "  Compressed: #{format_size(entry[:compressed_size])}"
         end
+        puts "  Modified: #{entry[:mtime]}" if entry[:mtime]
+        puts "  CRC32: 0x#{entry[:crc].to_s(16).upcase}" if entry[:crc]
       end
 
       def show_metadata(archive, pattern)
@@ -199,13 +200,7 @@ module Omnizip
       end
 
       def format_size(bytes)
-        return "0 B" if bytes.zero?
-
-        units = %w[B KB MB GB TB]
-        exp = (Math.log(bytes) / Math.log(1024)).to_i
-        exp = [exp, units.size - 1].min
-
-        "%.1f %s" % [bytes.to_f / (1024**exp), units[exp]]
+        Omnizip::CliOutputFormatter.format_size(bytes)
       end
     end
   end

data/lib/omnizip/commands/parity_create_command.rb

@@ -107,13 +107,7 @@ module Omnizip
       # @param bytes [Integer] Size in bytes
       # @return [String] Formatted size
       def format_size(bytes)
-        if bytes < 1024
-          "#{bytes} B"
-        elsif bytes < 1024 * 1024
-          "#{(bytes / 1024.0).round(1)} KB"
-        else
-          "#{(bytes / (1024.0 * 1024)).round(1)} MB"
-        end
+        Omnizip::CliOutputFormatter.format_size(bytes)
       end
     end
   end

data/lib/omnizip/extraction/selective_extractor.rb

@@ -33,7 +33,10 @@ module Omnizip
         FileUtils.mkdir_p(dest)
         extracted = []
 
-        entries_to_extract = list_matches
+        # Directory entries carry no content to extract
+        entries_to_extract = list_matches.reject do |entry|
+          entry_filename(entry).end_with?("/")
+        end
         total = entries_to_extract.size
         current = 0
 
@@ -162,7 +165,7 @@ module Omnizip
           entry.get_input_stream.read
         # allowed: some archives read entry content from the archive
         elsif @archive.respond_to?(:read)
-          @archive.read(entry)
+          @archive.read(entry_filename(entry))
         else
           raise Error, "Cannot read entry content"
         end
@@ -191,10 +194,19 @@ module Omnizip
 
       # Get filename from entry
       #
-      # @param entry [#entry_name] Entry (Omnizip::Entry or compatible)
+      # @param entry [#entry_name, #name] Entry (legacy Omnizip::Entry,
+      #   Archive::Entry from a reader session, or compatible)
       # @return [String] Filename
       def entry_filename(entry)
-        entry.entry_name || entry.to_s
+        # allowed: entries come from a caller-supplied archive
+        if entry.respond_to?(:entry_name)
+          entry.entry_name
+        # allowed: reader-session entries expose name, not entry_name
+        elsif entry.respond_to?(:name)
+          entry.name
+        else
+          entry.to_s
+        end
       end
 
       # Update progress tracker

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

@@ -64,7 +64,7 @@ module Omnizip
         # @raise [RuntimeError] if entry not found
         def extract_entry(entry_name, output_path)
           entry = @entries.find { |e| e.name == entry_name }
-          raise "Entry not found: #{entry_name}" unless entry
+          raise Errno::ENOENT, "Entry not found: #{entry_name}" unless entry
 
           extract_single_entry(entry, output_path)
         end

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

@@ -67,7 +67,7 @@ module Omnizip
         # @param output_path [String] Destination path
         def extract_entry(entry_path, output_path)
           entry = find_entry(entry_path)
-          raise "Entry not found: #{entry_path}" unless entry
+          raise Errno::ENOENT, "Entry not found: #{entry_path}" unless entry
 
           if entry.directory?
             FileUtils.mkdir_p(output_path)
@@ -111,7 +111,7 @@ module Omnizip
           loop do
             @io.seek(sector * Iso::SECTOR_SIZE)
             data = @io.read(Iso::SECTOR_SIZE)
-            raise "Failed to read volume descriptor" unless data
+            raise Omnizip::InvalidArchiveError, "Failed to read volume descriptor" unless data
 
             vd = VolumeDescriptor.parse(data)
 
@@ -127,7 +127,7 @@ module Omnizip
 
           return if @primary_volume_descriptor
 
-          raise "No primary volume descriptor found"
+          raise Omnizip::InvalidArchiveError, "No primary volume descriptor found"
         end
 
         # Parse directory structure

data/lib/omnizip/formats/iso/volume_descriptor.rb

@@ -29,7 +29,7 @@ module Omnizip
         #
         # @param data [String] Binary data
         def parse(data)
-          raise "Invalid volume descriptor size" unless data.bytesize >= 2048
+          raise Omnizip::InvalidArchiveError, "Invalid volume descriptor size" unless data.bytesize >= 2048
 
           # Byte 0: Volume descriptor type
           @type = data.getbyte(0)
@@ -37,7 +37,7 @@ module Omnizip
           # Bytes 1-5: Standard identifier "CD001"
           @identifier = data[1, 5]
           unless @identifier == ISO_IDENTIFIER
-            raise "Invalid ISO identifier: expected #{ISO_IDENTIFIER}, got #{@identifier}"
+            raise Omnizip::InvalidArchiveError, "Invalid ISO identifier: expected #{ISO_IDENTIFIER}, got #{@identifier}"
           end
 
           # Byte 6: Version (should be 1)

data/lib/omnizip/formats/rar/compression/ppmd/decoder.rb

@@ -230,26 +230,6 @@ module Omnizip
               # Decode as 8 direct bits
               @range_decoder.decode_direct_bits(8)
             end
-
-            # Decode RAR-specific escape code
-            #
-            # RAR variant H uses different escape code values
-            # and handling compared to standard PPMd7.
-            #
-            # Escape codes in RAR:
-            # - 0: New symbol follows
-            # - 1: Same as last symbol (run-length)
-            # - 2-255: Reserved for future use
-            #
-            # @return [Integer, nil] Escape code or nil
-            def decode_escape_code
-              # RAR escape codes differ from PPMd7
-              # This is a placeholder for the proper implementation
-
-              # For now, return 0 (new symbol follows)
-              # Real implementation would decode from range coder
-              0
-            end
           end
         end
       end