v1.4.0

• Fixed a data race on file fragments, causing issues when extracting archives with a lot of files.
• Re-wrote and tweaked the file extraction code to better use resources.
  • From my tests, you should see slightly low CPU usage and slightly higher memory usage, but with a slight performance increase. Importantly, performance seems to be more stable in general.
  • ExtractionOptions.SimultaneousFiles is depreciated and only ExtractionOptions.ExtractionRoutines is used.
• Nearly everywhere that binary.Read was used has been removed in favor of manually reading and decoding the bytes for better performance due to removing runtime reflection.
• squashfslow has has some slight API changes, mainly due to data.FullReader and data.Reader. I've formally decided that squashfslow's API is subject to change and added that to the squashfslow README.

What's Changed

• fix: prevent index out of range on long frag tables by @willmurphyscode in #30
• Redo extract by @CalebQ42 in #41

Full Changelog: v1.3.0...v1.4.0

Full Access

• Added OpenFile to get a *File instead of a fs.File
• Added access to squashfslow values for File and FS
• Re-use zstd & zlib decompressors.
  • I think the zstd implementation should be stable, but it took a while to get it working properly and I'm not sure exactly why it's working now 😛. I'll revert it if it ends up being unstable.
  • This should lower resource usage for zstd & zlib archives.

Performance!

• Significantly improved performance in nearly every instance.
  • Up to 4x improvements when using zstd.
  • Only minor improvements for the Arch Linux airootfs (xz compression).
  • Still significantly slower then unsquashfs (see README for more details)

Build tags

• Added the build tags no_gpl and no_obsolete to disable gpl libraries and obsolete compression type respectively.

What's Changed

• Allow mounting with an offset by @afbjorklund in #34
• Add list option to unsquashfs by @afbjorklund in #35
• Allow disabling lzo and lzma by @afbjorklund in #37

Full Changelog: v1.0.6...v1.1.0

.Mode() fixes

• Fixed missing file type flags when using .Mode(), most importantly for symlinks.

What's Changed

• Properly show symlinks in Mode by @afbjorklund in #33

New Contributors

• @afbjorklund made their first contribution in #33

Full Changelog: v1.0.5...v1.0.6

Reader fix

• Fixed issue caused by the data.Reader not emptying it's buffer on EOF causing it to replay the last block's data on subsequent Read calls (#32).

Fix stray println

What's Changed

• fix: remove stray println by @willmurphyscode in #31

New Contributors

• @willmurphyscode made their first contribution in #31

Full Changelog: v1.0.3...v1.0.4

Bug fix

Fixed issue with fragment, id, and indode reference table values on block borders reported by @willmurphyscode in #30.

Offset

Re-added NewReaderAtOffset

Improved performance

Improved extraction performance slightly by changing some pointers to normal values.