Highlights

This release switches to libbz2-rs-sys as the default bzip2 backend. It is written in rust, making it much easier to cross-compile. The new implementation is also more performant.

The rust implementation does not export C symbols by default. An explicit dependency on libbz2-rs-sys and setting the export-symbols feature flag will fix that if needed.

The bzip2-sys feature flag can be used to continue to use the C backend (either the system library of a C version built from source).

This is a breaking release, given that we want this update to be opt-in. We've used the occasion to perform an MSRV bump to 1.82.0. The libbz2-rs-sys crate has the bzip2-1.0.6 license. Arguably, that is the license that bzip2-sys should have had too, given that it is just a wrapper of the original bzip2 (and libbz2-rs-sys is a derivative work). You may need to allow this license when using cargo-deny.

What's Changed

• Release 0.6.0 by @folkertdev in #141

Full Changelog: v0.5.2...v0.6.0

Some minor fixes this time

• better wasm support (and we test wasm on CI)
• make bz_internal_error an extern fn by @folkertdev in #135 fixes a soundness issue. This is technically a semver-breaking change, but major versions for -sys crates create a lot of churn. We don't expect users to run into this

What's Changed

• fix: libc wasm types by @baszalmstra in #131
• update LICENSE-MIT by @rootdiae in #130
• run CI for wasm32-wasip2 by @folkertdev in #132
• remove libc dependency of bzip2-sys by @folkertdev in #133
• make bz_internal_error an extern fn by @folkertdev in #135

New Contributors

• @baszalmstra made their first contribution in #131
• @rootdiae made their first contribution in #130

Full Changelog: v0.5.1...v0.5.2

Highlights

Most changes relate to libbz2-rs-sys: we no longer enable the static feature, and now including different versions, that use either the C or the rust -sys crate, all work together in the same build.

What's Changed

• write::BzDecoder: Fix infinite loop on drop when no data is read or written by @chenxiaolong in #118
• don't enable the static feature when using libbz2-rs-sys by @folkertdev in #122
• update libbz2-rs-sys version that we use by @folkertdev in #123
• use core::ffi::{c_int, c_uint} in favor of libc by @folkertdev in #126
• Add finisher drop implementation to BzEncoder by @jonasbb in #121
• Update rand requirement from 0.8 to 0.9 by @dependabot in #127
• Disable bzip2-sys build script if rust backend is enabled by @NobodyXu in #125
• deprecate Compression::new by @folkertdev in #124
• build the docs on CI by @folkertdev in #128
• bump version to 0.5.1 and 0.1.12 by @folkertdev in #129

New Contributors

• @chenxiaolong made their first contribution in #118
• @jonasbb made their first contribution in #121
• @dependabot made their first contribution in #127
• @NobodyXu made their first contribution in #125

Full Changelog: v0.5.0...v0.5.1

The 0.5.0 release is the first release in ~2 years. It sets an msrv of 1.65.0 and pins the 2021 edition. If the msrv or edition are a problem for your project: just stick with the 0.4.4 release.

The repository has moved from @alexcrichton to @trifectatechfoundation, which takes on its maintainership from this point forward.

The main new feature of this release is the libbz2-rs-sys feature flag, which uses the pure rust https://github.com/trifectatechfoundation/libbzip2-rs implementation of bzip2. Being all rust, and using (by default) the rust allocator infrastructure, cross-compilation (e.g. to webassembly) should be much easier.

Besides that, the documentation was modernized (we can cross-link now) and several bugs were fixed.

What's Changed

• update partial_io to version 0.5.4 by @folkertdev in #108
• Remove tokio support by @folkertdev in #109
• support libbz2-rs-sys by @folkertdev in #110
• set msrv to 1.65.0 by @folkertdev in #111
• set the edition to 2021 by @folkertdev in #112
• fix clippy warnings by @folkertdev in #113
• update documentation by @folkertdev in #114
• Don't enable the stdio feature for the rust implementation of bzip2 by @bjorn3 in #117
• Fix infinite loop after invalid data by @jongiddy in #99
• prepare release 0.5.0 by @folkertdev in #115

New Contributors

• @folkertdev made their first contribution in #108
• @bjorn3 made their first contribution in #117
• @jongiddy made their first contribution in #99

Full Changelog: 0.4.4...v0.5.0