pulse v1.4.4

Summary

This release adds a local-only shutdown API for distributed tickers and includes several rmap/pool robustness improvements.

Highlights

• pool: add (*pool.Ticker).Close() to stop a node’s participation in a distributed ticker without deleting the shared ticker-map entry.
  • Use this for safe local restarts/handoff in multi-node pools.
  • Close() does not close the tick channel (matching time.Ticker semantics).

Full changelog (v1.4.3 → v1.4.4)

• 2f3aa04 pool: add Ticker.Close
• 6365d9f pool: cleanup maps on shutdown
• f9d8f4b rmap: add revisioning and list improvements
• 6037b55 Bump golang.org/x/net from 0.36.0 to 0.38.0 (#56)
• 28ed733 rmap: robustness + test Redis config (#63)

Compatibility

• Backwards compatible: existing Stop() behavior is unchanged (still deletes the shared ticker-map entry).

Links

• Compare: v1.4.3...v1.4.4

pulse v1.4.3

Fixes

• Fix rmap SetAndWait data race: remove the waiter notification channel close and make waiter notifications non-blocking, eliminating a race found by FAIL . [setup failed] in downstream consumers (notably goa-ai registry tests).
• Fix pool hashing data race: make the Jump Consistent Hash implementation concurrency-safe when routing/rebalancing worker events.

Notes

• No API changes; this is a bugfix release focused on race-detector correctness and stability under concurrency.

v1.4.2

What's Changed

• Fix potential job loss during concurrent worker cleanup/rebalancing under churn
• Add regression test TestJobLossDuringConcurrentWorkerCleanup to reproduce cascading failure scenario
• Fix weather example event unmarshalling to correctly error on truncated payloads

Full Changelog: v1.4.1...v1.4.2

v1.4.1

What's Changed

• Fix potential panics caused by an invalid regexp by returning an error by @douglaswth in #53
• Fix streaming Reader race condition by @douglaswth in #55

Full Changelog: v1.4.0...v1.4.1

v1.4.0

What's Changed

• pool: fix job payload cleanup during shutdown by @raphael in #52

Full Changelog: v1.3.0...v1.4.0

v1.3.0

What's Changed

• pool: Enhance worker lifecycle management and job reliability by @raphael in #51

Full Changelog: v1.2.0...v1.3.0

v1.2.0

What's Changed

• Enforce unique job keys: DispatchJob now returns an error when dispatching a job with a key that already exists
• Proper support for empty string values in rmap: setting an empty string value then calling Get now returns empty string and true.
• Update dependencies
• Adds SetIfNotExists to replicated maps that sets a value only if it isn't already in the map.
• Protect rmap.SetAndWait against corner cases

Full Changelog: v1.1.0...v1.2.0

v1.1.0

What's Changed

• Add back option to set reading block duration to node options by @nitinmohan87 in #46
• Add a debug log when worker fails to start a job by @raphael in #47
• Fix a few issues by @raphael in #48

New Contributors

• @nitinmohan87 made their first contribution in #46

Full Changelog: v1.0.6...v1.1.0

v1.0.6

What's Changed

• Fix rmap set debug log cardinality by @douglaswth in #17
• Reset timer when changing ticker duration by @raphael in #18
• Series of small fixes by @raphael in #20
• Stop workers before requeuing their jobs by @raphael in #22
• Test requeuing by @raphael in #21
• Fix issues with idle message requeuing by @raphael in #23
• Delete consumers when stale instead of reusing them. by @raphael in #24
• Add CI workflows by @raphael in #25
• Remove requeued jobs by @raphael in #26
• Improve logging by @raphael in #27
• Set read block duration to half worker TTL by @raphael in #29
• Remove requeued job from job map by @raphael in #28
• Make sure to expect requeue event ID by @raphael in #30
• Log Node ID with node map logs by @raphael in #31
• Log existing workers when joining by @raphael in #32
• Make logging more consistent by @raphael in #33
• Fix live lock with requeueJobs by @raphael in #34
• Periodically check for inactive workers by @raphael in #35
• Don't retry requeuing jobs with missing payloads by @raphael in #36
• Properly delete stale workers with no job by @raphael in #37
• Add more tests by @raphael in #38
• Add ability to list jobs and workers by @raphael in #39
• Make sure all timestamps are UTC by @raphael in #40
• Log panics occurring in event handler goroutines by @raphael in #41
• Timeout DispatchJob if no dispatch return is received by @raphael in #42
• Delete stale consumers not associated with sinks by @raphael in #43
• Reduce lock contention by @raphael in #44
• Fix lost job on requeue by @raphael in #45

Full Changelog: v0.0.6...v1.0.6

v0.0.6

What's Changed

• Ensure node.Close and node.RemoveWorker properly stop local jobs by @raphael in #15
• Update dependencies by @raphael in #16

Full Changelog: v0.0.5...v0.0.6