v0.6.4

What's Changed

Adds a fix for the raptor lake CPU bug and for an off-by-one error in adler32 on aarch64 resulting in an incorrect checksum.

• Run AVX-512 tests only if AVX-512 is available by @Shnatsel in #484
• Various CI improvements suggested by zizmor by @bjorn3 in #485
• run CI for powerpc64le by @folkertdev in #489
• use the vnni ilp trick for standard avx512 by @folkertdev in #488
• remove `dynamic-libz-sys by @folkertdev in #490
• prepare for miri supporting the crc intrinsics by @folkertdev in #492
• fix aliasing bug in infback by @folkertdev in #497
• clarify gzopen behavior around lseek64 by @folkertdev in #493
• run miri on aarch64 testing SIMD code by @folkertdev in #498
• fix internal API OOB read by @folkertdev in #496
• add uninit versions of stable compress and decompress by @folkertdev in #501
• clippy: unnecessary >= y + 1 or x - 1 >= by @folkertdev in #503
• document the behavior of window_bits by @folkertdev in #505
• [bug fix] adler32: reduce pair mod BASE after handle_tail on aarch64 NEON by @shenxiul in #504
• fix missing input validation in inflateBackInit_ by @folkertdev in #507
• return NULL when the requested allocation is too large by @folkertdev in #508
• fix and test aarch64_be by @folkertdev in #512
• disable aarch64_be miri run by @folkertdev in #517
• crc32: add options to ARM inline asm by @inahga in #516
• crc32: Add implementation for loongarch64 by @Gelbpunkt in #511
• bump libz-sys to 1.1.29 by @folkertdev in #519
• update Cargo.lock by @folkertdev in #521
• Work around crash on Intel Raptor Lake CPU by @glandium in #520
• test the loongarch64 crc32 implementation with Miri by @folkertdev in #523
• Use the stock zlib crc32 approach on aarch64 by @folkertdev in #513

New Contributors

• @Shnatsel made their first contribution in #484
• @shenxiul made their first contribution in #504
• @Gelbpunkt made their first contribution in #511
• @glandium made their first contribution in #520

Full Changelog: v0.6.3...v0.6.4