Replace the manual single-target ClientSession loops in the stream and TUI adapters with ManagedClient::start_with_subscription so the existing worker owns opening, deadline-aware pacing, receive handling, timeout polling, draining, and close.
Use bounded subscription waits for interruption and input polling, drain queued events before and after join, and report per-subscription drops without changing output, statistics, labels, summaries, warning, or peer-close policy. Add process and unit regressions for 10 ms cadence, interruption, final-event draining, TUI terminal state, and overflow visibility.
Advance delayed managed schedules to the latest due cadence slot and then directly to the first future deadline. Apply the same absolute-cadence behavior to the single runner and group burst/staggered pacing, while leaving the manual Client send API caller-paced.
Bound receive waits by the next send deadline so 10 ms intervals are not hidden behind the managed socket timeout. Add deterministic delayed-slot, finite-end, burst, staggered, and receive-bound regressions.
Snapshot sent counters when a runtime leaves its live session state, and expose PeerClosed from the single managed runner instead of folding it into TestComplete. Group timeout polling now recognizes a peer-closed runtime before attempting another operation, preserving the same provenance and counters.
Regression coverage verifies single and group peer-close outcomes, finite completion, cancellation, no-test completion, and final counters.
It was an attempt at handling integer wraparound more cleanly keeping in mind that we have a continuous mode,
and so we can more easily reach seq > u32::MAX
However after running the numbers, if we run with a interval of 10ms, it's still gonna take ~497 days before we hit integer overflow, so for now it doesn't seem worth to have this extra logic to handle something that'll in 99.9% of cases be purely an edge case
Keeping in mind autorate usage, in sqm-autorate-rust each send loop is 500 ms by default, and the sleep is then 500 ms / 5 reflectors, so our interval is 100 ms. that means it'll take 4970 days or ~13.6 years to reach the limit. Not worth dealing with IMO
Use internal monotonic time by default and introduce poll_timeouts_at for callers requiring a specific sampled Instant. This simplifies the API for standard usage while maintaining flexibility for external event loops.
Wrap the client execution in a cleanup handler to guarantee that event hub subscriptions are disconnected regardless of whether the session completes successfully or returns an error.
Nils Andreas Svee