An implementation from scratch of Zstandard in Wuffs. It is being developed following the published spec.

This is very much a work in progress, so it is not ready for use.

To build a simple zstd binary that will use stdin and stdout:

./build

And to test it with the input/output cases in testdata:

./test

Writing a decoder in Wuffs takes more time, but the end result is an implementation that is safe and can be used in many languages without linking against C.

For example, that would mean no cgo overhead with Go, and safer code for languages like Rust. Though that is somewhere in the future - see the roadmap.

If you're after a zstd implementation that works today, use https://github.com/DataDog/zstd.

This is the current progress of the decoder.

• Zstandard frames
  • Raw blocks
  • RLE blocks
  • Compressed blocks
    • Literals section
      • Raw literals block
      • RLE literals block
      • Compressed literals block
      • Treeless literals block
    • Sequences section
      • Predefined mode
      • RLE mode
      • Repeat mode
      • FSE compression mode
    • Sequence execution
      • Repeat offsets
      • Other offsets
  • XXH64 frame content checksum
• Skippable frames
• Dictionaries

These items are required for a stable 1.0 release:

• Wuffs 1.0 release
• Go support in Wuffs (generating a Go zstd library)
• Full zstd decoder implemented