1.21.1

1.21.1 (2024-01-08)

Bug Fixes

• patch: Fix incorrect public export of Lock (#217) (df284bb)

1.21.0

1.21.0 (2024-01-08)

Reduce dependencies from 12->7;

Removing extras-json and instead use Foundations JSON - we have a Foundation dependency anyway and will hopefully soon move to the new swift-foundation.

Fold in Progress.Swift inline with the project instead of depending on Ordo tagged release version

Fold in BenchmarkClock to remove dependency on Ordo package-datetime, move to use CLOCK_BOOTTIME

Update NIO locks from upstream (will be removed whenever we get locking primitives in the standard library...).

Move Benchmarks to a separate embedded project a la SwiftNIO to allow it to use newer toolchain and remove dependencies

Remove DocC plugin dependency as it is done behind the scenes by SPI anyway

What's Changed

• chore(minor): Reduce dependencies by @hassila in #215

Full Changelog: 1.20.0...1.21.0

1.20.0

1.20.0 (2023-12-21)

This release gives a significant performance boost for the benchmark result processing, baseline reading-parsing-exports etc, as the BenchmarkTool that is doing the heavy lifting now is built in release mode regardless of how the invocation of package benchmark was done.

Features

• minor: Manually build benchmarktool as release config (#214) (2e724bf)

1.19.0

1.19.0 (2023-12-20)

Bug Fixes

• minor: Fail CI also for update benchmark baseline updates that fails due to EPERM (#212) (2d3544e)

1.18.0

1.18.0 (2023-12-20)

Features

• minor: Make CI fail with regression for crashes or other run failures (#211) (645e475)

1.17.0

1.17.0 (2023-12-20)

Bug Fixes

• minor: Make CI fail for invalid arguments or failed benchmark builds (#210) (5fb6d7e)

1.16.0

1.16.0 (2023-12-15)

Features

• minor: Avoid rebuilds (#207) (05695c9)

1.15.0

1.15.0 (2023-12-12)

Features

Added support for the .microbenchmark metric set for convenience - also optimised the benchmark measurement overhead to make such microbenchmarks even better.

Fixed the rounding for results.

• minor: Performance optimization and a few bug fixes (#206) (81bfd1f)

1.14.0

1.14.0 (2023-12-11)

Features

Change the default output to be scaled to the inner loop count (and for throughput, always scale to this)

Add a new metric peakMemoryResidentDelta that discounts any test setup memory footprint and only shows the delta for the actual benchmark run for RSS.

What's Changed

Fixed TSAN failure
Fix (sometimes) cut off description
Proper checks for missing absolute threshold paths

• feat(minor): Various outstanding fixes by @hassila in #204

Full Changelog: 1.13.0...1.14.0

1.13.0

1.13.0 (2023-11-28)

Features

• minor: Lower required swift-system version to 1.1 to facilitate benchmark adoption for more users (#203) (f629323)