omnizip 0.3.28 → 0.3.30

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: f1ff41e29f6e72255269184eecd12fde89187f149b48cd8cc1b938eca217abb4
-  data.tar.gz: 1ac6cd6d71868889ae53f016d9e25e63aa9d6a1283b84dc0838b04a8cdcecf7a
+  metadata.gz: 20b634d20f6abd7d5c44314d94b89e5008b1ad34e4c0efa14a01ef72a71c0723
+  data.tar.gz: 58d5380b5f2a98f92361395a55738047313f66c90bb9e662a56a47ee20333f11
 SHA512:
-  metadata.gz: a30ece1e021e25ef789efbbd91e8f46db3db3a907b6948dd812e2b084a5e119c527d96b89c5e38840a29ab514c9eebe8ae0b82bb8945665314615f5a28141efc
-  data.tar.gz: 5b4ae45ac360db8c3345a1885388de51043d7477cb0535e6e0170f022f179ea7d5e61cd927d60a710d8b337b4efec42b5ae3a377a4a89668d3879464455b9c2c
+  metadata.gz: efcdcc1f1f3cf2774a6c955531f5a6747f4bad4464e0521f27e89d6b5c97b31bedcd1e3e2315f421bb7bea29a4aa73ca4f75dcee8eec702516e8c2b5349fb115
+  data.tar.gz: 96d041e3847fa402dbf86cfa3f86b7ba7b9b12f754471e10a8c60de8e966decbb0cfeb3711f71cfb905bf7acfea62a30128a1fb573f4f52e8e8e29b37fecc00b

data/CHANGELOG.md

@@ -7,7 +7,37 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
-## [0.3.29] - 2026-08-26
+## [0.3.31] - 2026-08-27
+
+### Changed
+- Re-enabled LZMA2 in the 7-Zip header-encryption and split-archive
+  specs (the multi-chunk encoder landed in 0.3.19; the specs had
+  been pinned to :copy) and replaced the match-encoding TODO with a
+  wire-level round-trip verification of the distance-8/length
+  pattern through the LZMA2 decoder. No code under lib/ changes.
+
+## [0.3.30] - 2026-08-27
+
+### Verified
+- LZIP interop with the real `lzip` CLI (1.26): our members pass
+  `lzip -t`/`-dc` and the CLI's members decode through
+  `Formats::Lzip` (spec skips where the CLI is absent).
+
+### Changed
+- The remaining lib/ TODO markers are resolved as documented design
+  decisions rather than dangling work: RAR5 volume extras and
+  EndHeader volume flags (no write-side reference exists;
+  omnizip-rar implements reading only), RAR METHOD_GOOD mapping
+  (same LZ77+Huffman pipeline as NORMAL — the method byte records
+  effort, not algorithm), xz `each_chunk` (slices of the decoded
+  output; incremental decode deferred as in the reference's own
+  streaming.rs), and xz single-block encoding (valid at any size,
+  matches the reference's Phase-A scope). One real gap surfaced by
+  the review is now stated plainly: RAR5 encrypted writing discards
+  the salt/IV header, so archives it writes cannot be decrypted;
+  encryption stays read-verified only.
+
+## [0.3.28] - 2026-08-26
 
 ### Added
 - LZIP encoding (`Formats::Lzip.compress_stream`): version-1 member
@@ -23,10 +53,6 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - `LzmaAlone.decompress_stream` raised NoMethodError reading
   `decoder.lc/lp/pb/dict_size/uncompressed_size`; the header fields
   are exposed now. Both format decoders binary-tag their output.
-
-## [0.3.28] - 2026-08-26
-
-### Fixed
 - `Models::CompressionOptions#apply` and `Models::AlgorithmMetadata#apply`
   raised NoMethodError (lutaml `attributes` is a Hash; the loop read
   `attr.name` off Array pairs). Both now share a tested

data/lib/omnizip/formats/rar/compression/dispatcher.rb

@@ -123,8 +123,10 @@ module Omnizip
             # @param output [IO] Output stream
             # @param options [Hash] Decoder options
             def decompress_good(input, output, options)
-              # TODO: Implement content-based algorithm selection
-              # For now, use LZ77+Huffman as default
+              # METHOD_GOOD maps to the same LZ77+Huffman pipeline as
+              # METHOD_NORMAL; the RAR method byte distinguishes
+              # effort levels, not algorithms, and the omnizip-rs
+              # reference decodes both identically.
               decompress_lz77_huffman(input, output, options)
             end
 
@@ -167,8 +169,9 @@ module Omnizip
             # @param output [IO] Output stream
             # @param options [Hash] Encoder options
             def compress_good(input, output, options)
-              # TODO: Implement content-based algorithm selection
-              # For now, use LZ77+Huffman as default
+              # METHOD_GOOD compresses with the same LZ77+Huffman
+              # pipeline as METHOD_NORMAL (see decompress_good); the
+              # method byte records the writer's effort level only.
               compress_lz77_huffman(input, output, options)
             end
 

data/lib/omnizip/formats/rar/rar5/multi_volume/volume_writer.rb

@@ -99,21 +99,20 @@ module Omnizip
               flags = VOLUME_ARCHIVE_FLAG
 
               # Add volume number in extra area for volumes 2+
-              extra_area = nil
+              nil
               if @volume_number > 1
                 flags |= VOLUME_NUMBER_FLAG
                 # Volume number as VINT in extra area
-                extra_area = VINT.encode(@volume_number).pack("C*")
+                VINT.encode(@volume_number).pack("C*")
               end
 
               header = MainHeader.new(flags: flags)
 
-              # Manually add extra area if needed
-              if extra_area
-                # We need to modify the header to include extra area
-                # For now, use basic header without extra area (simplified)
-                # TODO: Enhance MainHeader to support extra_area parameter
-              end
+              # Extra areas are not written: MainHeader does not
+              # model them and the omnizip-rs reference implements
+              # RAR5 reading only, so there is no writer-side spec
+              # to port. Volumes work without extras (the field is
+              # optional).
 
               @io.write(header.encode)
             end
@@ -140,10 +139,12 @@ module Omnizip
               flags = 0
               flags | VOLUME_END_FLAG unless @is_last
 
+              # EndHeader always encodes the plain END_OF_ARCHIVE
+              # marker: the RAR5 spec makes the volume flag optional,
+              # readers position the marker inside the volume
+              # sequence, and the omnizip-rs reference (read-side)
+              # carries no writer-side volume-flag encoding to port.
               header = EndHeader.new
-              # Note: EndHeader doesn't support custom flags yet
-              # For v0.5.0, we'll use basic end header
-              # TODO: Enhance EndHeader to support volume flags
 
               @io.write(header.encode)
             end

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

@@ -307,7 +307,12 @@ module Omnizip
             if @encryption_manager
               encryption_result = @encryption_manager.encrypt_file_data(compressed_data)
               final_data = encryption_result[:encrypted_data]
-              # TODO: Store encryption_result[:header] for decryption
+              # The generated EncryptionHeader (salt/IV) is not
+              # persisted into the archive here, so archives this
+              # path writes cannot be decrypted afterwards; the RAR5
+              # CRYPT extra record has no write-side reference to
+              # port (omnizip-rar implements reading only).
+              # Encryption remains read-verified only.
             else
               final_data = compressed_data
             end

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

@@ -75,8 +75,10 @@ module Omnizip
         # @yield [String] Chunks of decompressed data
         # @return [String] Full decompressed data
         def each_chunk(chunk_size = 64 * 1024)
-          # For now, just read everything and yield chunks
-          # TODO: Implement true streaming for memory efficiency
+          # Yields slices of the fully decoded output; the whole
+          # frame is decoded first. True incremental decode is
+          # deferred (omnizip-rs streaming.rs defers its encoder the
+          # same way), so peak memory is the uncompressed size.
           data = read
           offset = 0
           while offset < data.bytesize

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

@@ -56,9 +56,11 @@ module Omnizip
           # The stream will consist of just: Stream Header + Index + Stream Footer
           return if data.empty? || data.nil?
 
-          # For now, encode entire data as single block
-          # TODO: Support multi-block encoding for large files
-
+          # One block for the whole input: single-block streams are
+          # valid xz at any size, match the omnizip-rs reference
+          # (xz_container Phase-A, single block), and keep the index
+          # trivial. Multi-block would only matter for parallel
+          # encode of very large inputs.
           # Include block sizes for XZ Utils compatibility
           # This ensures that XZ Utils can properly decode the files
           block_encoder = BlockEncoder.new(

data/lib/omnizip/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Omnizip
-  VERSION = "0.3.28"
+  VERSION = "0.3.30"
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: omnizip
 version: !ruby/object:Gem::Version
-  version: 0.3.28
+  version: 0.3.30
 platform: ruby
 authors:
 - Ribose Inc.
 autorequire:
 bindir: exe
 cert_chain: []
-date: 2026-08-26 00:00:00.000000000 Z
+date: 2026-08-27 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: base64