omnizip 0.3.36 → 0.3.37

data/lib/omnizip/archive_handlers/zip_handler.rb

@@ -5,8 +5,13 @@ module Omnizip
     # Adapter that exposes the canonical +create+/+open+/+extract_to+
     # /+list+ interface for the ZIP format. Wraps +Omnizip::Zip::File+.
     class ZipHandler
-      def create(path, &block)
-        Omnizip::Zip::File.create(path, &block)
+      # Zip::File has no archive-level compression options (method and
+      # level are per-entry in Zip::OutputStream), so extra keywords —
+      # e.g. profile-injected ones — are absorbed here.
+      def create(path, **_options, &block)
+        Omnizip::Zip::File.create(path) do |file|
+          block&.call(FileProxy.new(file))
+        end
       end
 
       def open(path, &block)
@@ -66,6 +71,24 @@ module Omnizip
       def remove_entry(path, entry_name)
         Omnizip::Zip::File.open(path) { |zip| zip.remove(entry_name) }
       end
+
+      # Translates the generic +add(name, source_path)+ and
+      # +add_data(name, data)+ interface. +Zip::File#add_data+ is
+      # private, so in-memory content goes through the public block
+      # form +add(name) { data }+.
+      class FileProxy
+        def initialize(file)
+          @file = file
+        end
+
+        def add(name, source_path = nil)
+          @file.add(name, source_path)
+        end
+
+        def add_data(name, data)
+          @file.add(name) { data }
+        end
+      end
     end
   end
 end

data/lib/omnizip/archive_handlers.rb

@@ -8,5 +8,6 @@ module Omnizip
     autoload :ZipHandler, "omnizip/archive_handlers/zip_handler"
     autoload :TarHandler, "omnizip/archive_handlers/tar_handler"
     autoload :SevenZipHandler, "omnizip/archive_handlers/seven_zip_handler"
+    autoload :RarHandler, "omnizip/archive_handlers/rar_handler"
   end
 end

data/lib/omnizip/convenience.rb

@@ -40,13 +40,50 @@ module Omnizip
         return output_path
       end
 
-      Omnizip::ArchiveHandler.for(resolve_archive_format(output_path, format)).create(output_path) do |archive|
+      Omnizip::ArchiveHandler.for(resolve_archive_format(output_path, format)).create(output_path, **options) do |archive|
         archive.add(::File.basename(input_path), input_path)
       end
 
       output_path
     end
 
+    # Decompress a single-file compressed stream (the counterpart of
+    # #compress_file). The format is resolved from the input
+    # extension: .gz, .bz2, .xz, .lzma, .lz.
+    #
+    # @param compressed_path [String] Path to the compressed file
+    # @param output_path [String] Path to write the decompressed data
+    # @return [String] +output_path+
+    def decompress_file(compressed_path, output_path)
+      require_archive!(compressed_path)
+
+      format_class = DECOMPRESSORS[::File.extname(compressed_path).downcase]
+      unless format_class
+        raise Omnizip::UnsupportedFormatError,
+              "decompress_file supports #{DECOMPRESSORS.keys.join(', ')}; " \
+              "use extract_archive for archive formats"
+      end
+
+      case format_class
+      when :xz
+        ::File.binwrite(output_path, Formats::Xz.decompress(compressed_path))
+      when :zstandard
+        zstd = Algorithms::Zstandard.new
+        ::File.open(compressed_path, "rb") do |input_io|
+          ::File.open(output_path, "wb") do |output_io|
+            zstd.decompress(input_io, output_io)
+          end
+        end
+      else
+        ::File.open(compressed_path, "rb") do |input_io|
+          ::File.open(output_path, "wb") do |output_io|
+            format_class.decompress_stream(input_io, output_io)
+          end
+        end
+      end
+      output_path
+    end
+
     # Compress a directory into an archive.
     #
     # @param input_dir [String] Path to input directory
@@ -64,12 +101,23 @@ module Omnizip
         raise ArgumentError, "Input is not a directory: #{input_dir}"
       end
 
+      # A directory cannot be written through a single-file stream
+      # format; routing it to the default would silently produce a
+      # ZIP under a foreign extension.
+      if format == DEFAULT_FORMAT && single_file_handler(output_path)
+        raise Omnizip::UnsupportedFormatError,
+              "#{::File.extname(output_path).downcase} is a single-file " \
+              "stream format and cannot contain a directory; use an " \
+              "archive format (.zip/.tar/.7z) or compress entries " \
+              "individually"
+      end
+
       if options[:profile]
         first_file = find_first_file(input_dir)
         apply_profile(first_file, options)
       end
 
-      Omnizip::ArchiveHandler.for(resolve_archive_format(output_path, format)).create(output_path) do |archive|
+      Omnizip::ArchiveHandler.for(resolve_archive_format(output_path, format)).create(output_path, **options) do |archive|
         add_directory_contents(archive, input_dir, "", recursive: recursive)
       end
 
@@ -87,7 +135,7 @@ module Omnizip
     def extract_archive(archive_path, output_dir, format: DEFAULT_FORMAT,
                         overwrite: false, **options)
       require_archive!(archive_path)
-      Omnizip::ArchiveHandler.for(resolve_archive_format(archive_path, format))
+      Omnizip::ArchiveHandler.for(resolve_archive_format(archive_path, format, writing: false))
         .extract_to(archive_path, output_dir,
                     overwrite: overwrite, **options)
     end
@@ -102,7 +150,7 @@ module Omnizip
     def list_archive(archive_path, format: DEFAULT_FORMAT, details: false,
                      **options)
       require_archive!(archive_path)
-      Omnizip::ArchiveHandler.for(resolve_archive_format(archive_path, format))
+      Omnizip::ArchiveHandler.for(resolve_archive_format(archive_path, format, writing: false))
         .list(archive_path, details: details, **options)
     end
 
@@ -114,7 +162,7 @@ module Omnizip
     # @return [String] Entry contents
     def read_from_archive(archive_path, entry_name, format: DEFAULT_FORMAT)
       require_archive!(archive_path)
-      Omnizip::ArchiveHandler.for(resolve_archive_format(archive_path, format)).read_entry(archive_path, entry_name)
+      Omnizip::ArchiveHandler.for(resolve_archive_format(archive_path, format, writing: false)).read_entry(archive_path, entry_name)
     end
 
     # Add a file to an existing archive.
@@ -188,6 +236,26 @@ module Omnizip
     # honest behavior.
     READ_ONLY_FORMAT_EXTENSIONS = [".rar", ".iso", ".cpio"].freeze
 
+    # Extensions whose format has a READ-ONLY handler: extraction and
+    # listing route to it, while creation keeps raising.
+    READ_ARCHIVE_FORMAT_EXTENSIONS = {
+      ".rar" => :rar,
+    }.freeze
+
+    # Extension -> single-file decompressor (stream interface).
+    # The Gzip/Bzip2File/Xz classes take path or stream; use the
+    # ones exposing decompress_stream for symmetry.
+    # Xz exposes a path-based decompress rather than a stream one,
+    # so it is marked specially here.
+    DECOMPRESSORS = {
+      ".gz" => Formats::Gzip,
+      ".bz2" => Formats::Bzip2File,
+      ".xz" => :xz,
+      ".lzma" => Formats::LzmaAlone,
+      ".lz" => Formats::Lzip,
+      ".zst" => :zstandard,
+    }.freeze
+
     SINGLE_FILE_COMPRESSORS = {
       ".gz" => lambda do |input, output, options|
         Omnizip::Formats::Gzip.compress(input, output, options)
@@ -214,6 +282,14 @@ module Omnizip
           end
         end
       end,
+      ".zst" => lambda do |input, output, options|
+        ::File.open(input, "rb") do |input_io|
+          ::File.open(output, "wb") do |output_io|
+            Omnizip::Algorithms::Zstandard.new.compress(input_io, output_io,
+                                                        options)
+          end
+        end
+      end,
     }.freeze
 
     # The single-file compressor matching the output extension, or nil.
@@ -225,12 +301,14 @@ module Omnizip
     # Archive format for a path: an explicit +format+ wins unless it
     # is the default; otherwise the extension routes. Known-but-
     # unwritable extensions raise instead of receiving a mislabeled
-    # ZIP.
-    def resolve_archive_format(path, format)
+    # ZIP. +writing:+ false (read operations) additionally routes
+    # extensions with read-only handlers.
+    def resolve_archive_format(path, format, writing: true)
       return format unless format == DEFAULT_FORMAT
 
       ext = ::File.extname(path).downcase
       routed = ARCHIVE_FORMAT_EXTENSIONS[ext]
+      routed ||= READ_ARCHIVE_FORMAT_EXTENSIONS[ext] unless writing
       return routed if routed
 
       if READ_ONLY_FORMAT_EXTENSIONS.include?(ext)

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

@@ -121,6 +121,10 @@ module Omnizip
 
           header_flags = read_vint(io)
 
+          # Optional ExtraAreaSize / DataSize precede the file fields
+          extra_size = header_flags.anybits?(RAR5_FLAG_EXTRA_AREA) ? read_vint(io) : 0
+          data_size = header_flags.anybits?(RAR5_FLAG_DATA_AREA) ? read_vint(io) : 0
+
           # Read file header
           file_flags = read_vint(io)
           unpack_size = read_vint(io)
@@ -145,7 +149,7 @@ module Omnizip
 
           # Set entry properties
           entry.size = unpack_size
-          entry.compressed_size = 0 # Not directly available in header
+          entry.compressed_size = data_size
           entry.crc = crc
           entry.host_os = host_os
           entry.flags = file_flags
@@ -154,17 +158,9 @@ module Omnizip
           entry.is_dir = file_flags.anybits?(RAR5_FLAG_IS_DIR)
           entry.version = 5
 
-          # Read extra area if present
-          if header_flags.anybits?(RAR5_FLAG_EXTRA_AREA)
-            extra_size = read_vint(io)
-            io.read(extra_size) if extra_size.positive?
-          end
-
-          # Read data area if present
-          if header_flags.anybits?(RAR5_FLAG_DATA_AREA)
-            data_size = read_vint(io)
-            io.read(data_size) if data_size.positive?
-          end
+          # Skip extra area, then compressed data
+          io.read(extra_size) if extra_size.positive?
+          io.read(data_size) if data_size.positive?
 
           entry
         end

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

@@ -58,7 +58,9 @@ module Omnizip
         RAR5_FLAG_UNKNOWN_BLOCKS = 0x0004
         RAR5_FLAG_DATA_INHERITED = 0x0008
         RAR5_FLAG_CHILD_BLOCKS = 0x0010
-        RAR5_FLAG_IS_DIR = 0x0020
+        # File-header-specific flags (FileFlags field): the directory
+        # bit is 0x0001 per the RAR5 spec.
+        RAR5_FLAG_IS_DIR = 0x0001
         RAR5_FLAG_MULTI_VOLUME = 0x0001
 
         # Compression methods

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

@@ -122,9 +122,11 @@ module Omnizip
         #
         # @param io [IO] Input stream
         def parse_rar5_header(io)
-          # RAR5 uses variable-length integer encoding
+          # RAR5 uses variable-length integer encoding. Layout per the
+          # RAR 5.0 format: HeaderCRC32(4), Header size(vint), Header
+          # type(vint), Header flags(vint), then type-specific fields.
           read_uint32(io)
-          read_vint(io)
+          read_vint(io) # header size (redundant: areas carry their own)
           header_type = read_vint(io)
           header_flags = read_vint(io)
 
@@ -133,12 +135,20 @@ module Omnizip
           end
 
           @flags = header_flags
-          @is_multi_volume = header_flags.anybits?(RAR5_FLAG_MULTI_VOLUME)
 
-          # Read extra area if present
-          return unless header_flags.anybits?(RAR5_FLAG_EXTRA_AREA)
+          extra_size = 0
+          if header_flags.anybits?(RAR5_FLAG_EXTRA_AREA)
+            extra_size = read_vint(io)
+          end
+
+          # Archive flags: 0x0001 volume, 0x0002 volume number
+          # present, 0x0004 solid, 0x0008 recovery, 0x0010 locked
+          archive_flags = read_vint(io)
+          @is_multi_volume = archive_flags.anybits?(0x0001)
+          @is_solid = archive_flags.anybits?(0x0004)
+          @is_locked = archive_flags.anybits?(0x0010)
+          read_vint(io) if archive_flags.anybits?(0x0002) # volume number
 
-          extra_size = read_vint(io)
           io.read(extra_size) if extra_size.positive?
         end
 

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

@@ -142,8 +142,12 @@ module Omnizip
           # Detect recovery records
           detect_recovery_records
 
-          # Parse entries using decompressor
-          parse_entries_with_decompressor
+          # Parse entries natively first (full metadata: sizes,
+          # directory flags); fall back to the external decompressor
+          # when native parsing yields nothing (e.g. encrypted
+          # headers).
+          parse_entries_from_header
+          parse_entries_with_decompressor if @entries.empty?
         end
 
         # Detect recovery records in archive

data/lib/omnizip/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Omnizip
-  VERSION = "0.3.36"
+  VERSION = "0.3.37"
 end

data/lib/omnizip.rb

@@ -83,6 +83,7 @@ module Omnizip
   autoload :Implementations, "omnizip/implementations"
   autoload :ArchiveHandler, "omnizip/archive_handler"
   autoload :ArchiveHandlers, "omnizip/archive_handlers"
+  autoload :Archive, "omnizip/archive"
   autoload :Password, "omnizip/password"
   autoload :Convenience, "omnizip/convenience"
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: omnizip
 version: !ruby/object:Gem::Version
-  version: 0.3.36
+  version: 0.3.37
 platform: ruby
 authors:
 - Ribose Inc.
@@ -319,8 +319,12 @@ files:
 - lib/omnizip/algorithms/zstandard/sequences.rb
 - lib/omnizip/algorithms/zstandard/sequences_encoder.rb
 - lib/omnizip/algorithms/zstandard/xxhash.rb
+- lib/omnizip/archive.rb
+- lib/omnizip/archive/builder.rb
+- lib/omnizip/archive/reader_session.rb
 - lib/omnizip/archive_handler.rb
 - lib/omnizip/archive_handlers.rb
+- lib/omnizip/archive_handlers/rar_handler.rb
 - lib/omnizip/archive_handlers/seven_zip_handler.rb
 - lib/omnizip/archive_handlers/tar_handler.rb
 - lib/omnizip/archive_handlers/zip_handler.rb