v0.6.0

This release adds support for multithreading, and enables dithering of u8 output.

What's Changed

• Do not use nightly rust-analyzer for Nix dev shells by @tirr-c in #843
• Add missing dependencies for delta palette. by @veluca93 in #846
• Add some missing Send+Sync bounds for MT by @veluca93 in #844
• Fix group border detection in render pipeline by @veluca93 in #845
• Prepare for multithreading support by @veluca93 in #849
• Improve handling of input buffers in the render pipeline by @veluca93 in #850
• Make the render pipeline interface MT-friendly by @veluca93 in #851
• Use mimalloc as the allocator for jxl_cli. by @veluca93 in #855
• Prepare Modular and VarDCT decoding for multithreading. by @veluca93 in #852
• Port blue noise dithering from libjxl for 8-bit output by @jonnyawsom3 in #841
• Grow section buffers from available input, not TOC-declared size by @hjanuschka in #856
• Parse Mul clamp without extra channels by @hjanuschka in #858
• Initial support for parallel decoding by @veluca93 in #853
• Implement parallel decoding of LF groups. by @veluca93 in #854
• Remove some allow(unsafe_code), and make a T::cast_slice helper. by @veluca93 in #860
• Don't force eager rendering when partial rendering is disallowed by @hjanuschka in #863
• Continue refilling after incremental parser progress by @hjanuschka in #866
• Replace AtomicRefCell with Mutex/RwLock. by @veluca93 in #864
• Avoid reentrant read lock on aliased squeeze average buffer by @hjanuschka in #867
• Fill vertically subsampled border rows with floor semantics by @hjanuschka in #868
• Take read locks on palette prediction context grids by @hjanuschka in #869
• Use deterministic iteration order for decode scheduling sets by @hjanuschka in #870
• Fix smooth-unsqueeze scratch row loading races by @hjanuschka in #871
• Use shuttle for concurrency testing. by @veluca93 in #862
• Clip low-memory ready rectangles to the input size by @hjanuschka in #873
• Count partial groups when limiting render parallelism by @tnikkel in #872
• Fix rendering of subsampled frames by @tirr-c in #875
• Allow subsampled LF smoothing by @jonnyawsom3 in #861
• Change a Relaxed fetch-add into an AcqRel. by @veluca93 in #876
• Remove unused option progressive_mode by @sboukortt in #880
• update README by @mo271 in #879
• v0.6.0 by @veluca93 in #874

Full Changelog: v0.5.1...v0.6.0