2.3.1

What’s Changed

• Configure maven-gpg (#458) @pivovarit
• Describe params (#457) @pivovarit
• Describe params (#456) @pivovarit
• Remove unsupported JavaDocs (#455) @pivovarit
• Make ParallelCollectors.Batching public (#454) @pivovarit
• Refactor ArchitectureTest (#453) @pivovarit

Authors

@pivovarit

2.3.0

What’s Changed

• Fix ordered batching stream not maintaining the order (#449) @pivovarit
• Spliterator-based BatchingStream (#444) @pivovarit
• Optimize batching (#443) @pivovarit
• Optimize BatchingStream for single batches (#442) @pivovarit
• Hide internal method (#440) @pivovarit
• Improve performance of parallelism == 1 (#436) @pivovarit
• Remove unnecessary batching infrastructure (#435) @pivovarit
• Pre-size batching accumulator (#434) @pivovarit
• Do not expose values wrapped as CompletableFuture as finisher parameters (#433) @pivovarit
• Pre-size batching accumulator (#432) @pivovarit
• Reactive exception handling and cancellation (#430) @pivovarit
• Add ParallelCollectors#toFuture (#429) @pivovarit
• Remove redundant InFlight class (#425) @pivovarit

Authors

@dependabot-preview, @dependabot-preview[bot] and @pivovarit

2.2.0

What’s Changed

• Collect to list directly without allocating intermediate and short-lived objects (#422) @pivovarit
• Optimize for parallelism of one (#418) @pivovarit
• Add comparisons with streaming collectors (#413) @pivovarit
• Add basic benchmarks (#409) @pivovarit
• Don't allocate objects when processing batched objects (#408) @pivovarit
• Create a Spliterator of known size when batching (#402) @pivovarit

Authors

@pivovarit

2.1.0

Introduced a new subset of drop-in replacement collectors ParallelCollectors.Batching.

These implement a different task-splitting strategy where operations are performed in batches and not separately (one object in a thread pool's worker queue represents a batch of operations to be performed by a single thread):

ParallelCollectors.Batching:

• parallel(...)
• parallelToStream(...)
• parallelToOrderedStream(...)

2.0.0

Redesigned and simplified the public-facing API to enable reuse of java.util.stream.Collector instances for providing custom reduction strategies which enabled reduction of the public API from 13 methods to just 4:

• CompletableFuture<List<T>> parallel(Function, Collector, Executor, parallelism)

• CompletableFuture<Stream<T>> parallel(Function, Executor, parallelism)

• Stream<T> parallelToStream(Function, Executor, parallelism)

• Stream<T> parallelToOrderedStream(Function, Executor, parallelism)

1.2.1

• adjusted visibility modifiers of some API methods

1.2.0

• added new parallel methods that are interoperable with java.util.stream.Collectors
• simplified internal implementations
• deprecated parallelTo(List/Set/Map/Collection) methods in favor of the above one

1.1.0

• introduce interruptions of tasks when an exception gets thrown
• deprecate parallelMap and parallelMapOrdered and replace them with parallel and parallelOrdered

1.0.3

• Reimplemented CompletionOrderSpliterator with throughput in mind

1.0.2

• fixed improperly configured CompletionStreamSpliterator