v0.1.2

What's Changed

Mostly performance improvements, primarily to compression. We've also added the semver-prefix feature flag, that will prefix all exposed symbols with a version-specific prefix, so that we don't clash with the C library (or different versions of ourselves). That means multiple versions of bzip2 can be used in the same dependency tree without any weird behavior.

• run the bzip2-rs test suite by @folkertdev in #70
• Add NLnet funding to readme by @erikjee in #71
• refactors in un_rle_obuf_to_output_fast by @folkertdev in #72
• generate the crc32 lookup table at const time by @folkertdev in #73
• add an example for compression (for benchmarking) by @folkertdev in #74
• extract the save area by @folkertdev in #75
• Decompress shrink more fields by @folkertdev in #76
• Remove double bounds check by @folkertdev in #77
• Shrink huffman decode tables by @folkertdev in #78
• Remove global fields by @folkertdev in #79
• Bzlib macros by @folkertdev in #80
• cleanups by @folkertdev in #81
• Speedup compress by @folkertdev in #83
• Add CI benchmarking by @bjorn3 in #82
• decompress cleanups by @folkertdev in #84
• Update the readme by @bjorn3 in #86
• Revert accidental change in previous PR by @bjorn3 in #87
• decompress.rs: faster input reading by @folkertdev in #85
• Cleanup blocksort more by @folkertdev in #89
• add semver-prefix feature for use by bzip2-rs by @folkertdev in #92
• read coding tables in one go if there is sufficient input by @folkertdev in #90

New Contributors

• @erikjee made their first contribution in #71

Full Changelog: v0.1.1...v0.1.2