v0.6.0

This release updates the Decoder API to decode header fields incrementally rather than emitting them all at once (#67). This cuts allocations, improves performance, and enables consumers to process fields as they are decoded.

What's Changed

• ci: update golangci-lint to v2, run it on both Go versions by @marten-seemann in #57
• bump Go version to 1.23 in go.mod, update CI config by @marten-seemann in #56
• update to Go 1.25, drop Go 1.23 by @marten-seemann in #59
• Bump actions/setup-go from 5 to 6 by @dependabot[bot] in #60
• Bump actions/checkout from 4 to 5 by @dependabot[bot] in #58
• ci: remove 386 (32 bit x86) by @marten-seemann in #61
• ci: enable ClusterFuzzLite PR fuzzing by @marten-seemann in #62
• Bump golangci/golangci-lint-action from 8 to 9 by @dependabot[bot] in #63
• ci: fix Codecov configuration by @marten-seemann in #65
• add benchmarks for the decoder by @marten-seemann in #64
• ci: disable fail-fast for all matrix jobs by @marten-seemann in #68
• example: fix input file parsing by @marten-seemann in #69
• interop: update qifs to da52cd9 by @marten-seemann in #70
• simplify testing setup by @marten-seemann in #71
• rework Decoder API to decode header fields one by one by @marten-seemann in #67
• use io.ErrUnexpectedEOF error when input is too short by @marten-seemann in #73
• README: improve description and fix interop command by @marten-seemann in #74

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