Fluxzero 1.151.1

1.151.1 (2026-04-14)

Bug Fixes

• fix(tracking): remove unused join in pause event publishing (3e4ab3f)
• fix(tracking): preserve interrupts while storing position (763c50d)

  Handle InterruptedException explicitly when waiting for tracker position updates so storePosition does not lose the thread interrupt state or keep retrying it as a normal failure.

  This keeps tracker shutdown and cancellation responsive and adds focused test coverage for interrupted position storage.

• fix(websocket): shut down result handling on close (29b9dbd)

  Shut down the websocket result executor when the client closes so background result handling threads do not outlive the client lifecycle.

  This keeps websocket executor cleanup consistent with the ping and reconnect executors while preserving the existing websocket test coverage.

• fix(websocket): run ping timeouts on a worker pool (6364e57)

  Configure the websocket ping scheduler with a dedicated worker pool so expired ping tasks no longer execute inline on the scheduler thread.

  This keeps ping timeout handling from stalling on a single scheduler thread and adds focused tests for async task delegation and websocket ping scheduler configuration.

• fix(websocket): make reconnect retries easier to follow (1ac3f5e)

  Log repeated websocket connection failures more frequently and include retry counts in both retry and success messages.

  This makes long-running reconnect loops easier to diagnose in production and adds focused tests around the retry logging cadence.

• fix(websocket): configure connection timeouts explicitly (305fc9b)

  Add an explicit websocket connection timeout to the client config, propagate it to the underlying websocket container, and wrap connection attempts in a slightly longer outer failsafe.

  This makes connection setup behavior more predictable, keeps the public config independent from a specific container implementation, and adds focused tests for timeout propagation and outer timeout cancellation.

• fix(websocket): validate session pool size upfront (88a2470)

  Reject websocket session pools smaller than one and simplify round-robin index updates with floorMod for clearer, safer slot selection.

  This prevents invalid pool configurations from failing later at runtime and adds focused tests for zero, negative, and single-session pool behavior.

• fix(websocket): retry requests off the close thread (24e030c)

  Schedule outstanding websocket request retries on a dedicated reconnect executor instead of running them inline from onClose or abort-driven close handling.

  This keeps reconnect work away from timeboxed close threads, avoids interruption-driven reconnect failures, and adds focused test coverage for backlog-gated async retries without timing-sensitive assertions.

• fix(retries): stop retry loops on thread interruption (0b91c99)

  Treat thread interruption as a cancellation signal inside retryOnFailure instead of retrying it as a normal failure.

  This prevents reconnect and other retry-driven flows from silently restarting after an interrupted task and makes the retry behavior easier to reason about.