Observed failing intermittently in Release CI (Build & Test job) but not in Debug, on PR #595 (unrelated subsystem — that PR only touches B3.Exchange.Gateway FIXP session/journal code, not B3.Exchange.Persistence's Channel Dispatcher).
Failure:
Assert.Equal() Failure: Values differ
Expected: 2
Actual: 0
at B3.Exchange.Persistence.Tests.ChannelDispatcherStartupEpochTests.RestoreReplay_StartsDurableEpoch_PreservesState_AndOrdersResetSnapshotLive() line 356
Run: https://github.com/pedrosakuma/B3MatchingPlatform/actions/runs/31147435671/job/92769710797
Rerun of the same commit (Build & Test job only) passed cleanly: https://github.com/pedrosakuma/B3MatchingPlatform/actions/runs/31147435671/job/92850357903
Only failed in Release configuration, not Debug — suggests a timing/ordering race exposed by JIT optimization (e.g., a WAL replay/reset-publication race under RestoreReplay), not a genuine regression from #595's diff. Filing per repo flaky-test tracking convention so future occurrences can be matched against this signature instead of re-investigated from scratch.
Observed failing intermittently in Release CI (
Build & Testjob) but not in Debug, on PR #595 (unrelated subsystem — that PR only touchesB3.Exchange.GatewayFIXP session/journal code, notB3.Exchange.Persistence's Channel Dispatcher).Failure:
Run: https://github.com/pedrosakuma/B3MatchingPlatform/actions/runs/31147435671/job/92769710797
Rerun of the same commit (
Build & Testjob only) passed cleanly: https://github.com/pedrosakuma/B3MatchingPlatform/actions/runs/31147435671/job/92850357903Only failed in Release configuration, not Debug — suggests a timing/ordering race exposed by JIT optimization (e.g., a WAL replay/reset-publication race under
RestoreReplay), not a genuine regression from #595's diff. Filing per repo flaky-test tracking convention so future occurrences can be matched against this signature instead of re-investigated from scratch.