Fragments work properly

Fixed a couple issues with fragments. Particularly:

• If you had more then 512 fragments then it would start using ID table numbers (oops) (thanks @tri-adam)
• Some fragments would not be read from deep enough into the fragment

Additional Fixes:

• FS.ReadDir, FS.Glob, and FS.Stat would throw type assertion errors. (Again, thanks @tri-adam)

What's Changed

• v0.6.0 fixes by @tri-adam in #16

Full Changelog: v0.6.0...v0.6.1

Mainly just some code work

Ideally this release does basically nothing, it's just a large re-write of many parts of the code to make it more maintainable and understandable. This does fix one bug (#15 and #14) relating to how fragments are read and processed.

There are some breaking changes.

• NewSquashfsReader and NewSquashfsReaderFromReader have been renamed to NewReader and NewReaderFromReader.
  • This better matches other libraries.
• Reader.Init() was removed. This was never meant to be exported and shouldn't have been used.
• FileInfo and DirEntry is not longer exported. They were identical in functionality to fs.FileInfo and fs.DirEntry so it felt unnecessary to export them as separate structs.

This release should be good, but due to the large amount of changes (and the fact I apparently finished it a month ago, then go COVID and completely forgot about it) I recommend caution when upgrading to it.

Handle paths with special characters

Changed Open, ReadDir, Stat, and Sub to only accept exact matches. Allowing wildcards by using path.Match causes issues in certain circumstances. You can use fs.Glob if you need to use wildcards, and I'll probably implement fs.GlobFS at some point in the near (ish) future.

What's Changed

• fix: handle paths with special characters by @tri-adam in #11

New Contributors

• @tri-adam made their first contribution in #11

Full Changelog: v0.5.3...v0.5.4

Added "." directory support.

Added support for "." as a directory thanks to @stffabi

Fixed io.Reader support

Fixed some issues I was having with io.Readers and squashfs archives

io.Reader support

You can now use an io.Reader to create a squashfs.Reader. This is not recommended, especially for large archives since almost all of the archive WILL end up cached in memory.

Lzo and Xz filters

• This library now has support for Lzo compression and Xz filters.

io/fs

This library now implements several io/fs interfaces, including fs.FS and fs.File for the main Reader and File respectively. In making it compliant, I did manage to slightly speed up extraction slightly.

Hopefully, the next release will be v1.0.0 when I finish Writer.

Even MORE BUGS?

Fixed extended inode parsing

More bugs?

Fixed a bug relating to extended symlink inodes.