feat(clients): CLI-15 surface ReplayGap reconnect sentinel as typed signal (4/5 clients)

The gateway emits a ReplayGap sentinel MxEvent at the head of a StreamEvents
stream resumed via after_worker_sequence when the requested cursor predates the
oldest retained event. Clients previously ignored it, silently mis-treating a
lossy resume as continuous. Each client now surfaces the sentinel as a distinct,
typed, non-terminal signal (never synthesized, never swallowed) so a consumer can
detect the gap and re-snapshot; resume contract is
after_worker_sequence = oldest_available_sequence - 1.

- .NET: MxEventStreamItem (IsReplayGap/ReplayGap/Event) via new StreamEventItemsAsync
  + AsStreamItemsAsync extension. Build clean, 87 passed.
- Go: EventResult.ReplayGap field + IsReplayGap(); ReplayGap type alias. build/vet/test clean.
- Rust: EventItem enum (Event/ReplayGap); EventStream now yields Result<EventItem, Error>;
  CLI renders REPLAY_GAP line / replayGap JSON. fmt/check/test/clippy clean.
- Python: ReplayGap dataclass; stream_events yields pb.MxEvent | ReplayGap. 131 passed.
- Shared docs: ClientLibrariesDesign non-goals reframed (reconnect-replay protocol is
  consumable; auto-reconnect stays a non-goal); CrossLanguageSmokeMatrix resume-gap note.

Java client is deferred to the windev batch (no local JRE); CLI-15 stays open until it lands.

Claude-Session: https://claude.ai/code/session_01DMXXvNuPekkkrTEyPNxEkW
This commit is contained in:
Joseph Doherty
2026-07-09 16:22:28 -04:00
parent c823a7b60b
commit 0c6e5b3ace
22 changed files with 972 additions and 37 deletions
+42
View File
@@ -84,6 +84,48 @@ messages. `MxGatewaySession.OpenSessionReply` keeps the raw session-open reply
available, and command helpers have `*RawAsync` variants when callers need the
complete `MxCommandReply`.
### Event Streaming And Replay Gaps
`StreamEventsAsync(afterWorkerSequence)` yields raw generated `MxEvent`
messages. Passing a non-zero `afterWorkerSequence` resumes a session's event
stream after a known worker sequence — this is the reconnect cursor. If that
cursor is *stale* — older than the oldest event the gateway still retains in the
session replay ring — the events in between were evicted and cannot be replayed.
The gateway signals this by emitting a single **replay-gap sentinel** at the head
of the resumed stream: an `MxEvent` with its `ReplayGap` field set, `Family`
unspecified, and no body. It means "you missed events — discard local state and
re-snapshot."
Rather than force callers to inspect the raw sentinel, the client exposes a
typed surface, `StreamEventItemsAsync`, which yields `MxEventStreamItem` values:
```csharp
await foreach (MxEventStreamItem item in session.StreamEventItemsAsync(
afterWorkerSequence: lastSeenSequence))
{
if (item.IsReplayGap)
{
// We missed events: throw away local state and re-snapshot.
ReplayGap gap = item.ReplayGap!;
// Resume without incurring another gap:
lastSeenSequence = gap.OldestAvailableSequence - 1;
await ReSnapshotAsync();
continue;
}
HandleEvent(item.Event); // normal MXAccess event, IsReplayGap == false
lastSeenSequence = item.Event.WorkerSequence;
}
```
The typed surface never synthesizes or drops events — it only makes the
gateway's own sentinel observable. Normal events pass through with
`IsReplayGap == false` and `ReplayGap == null`. The gap is only ever produced by
`StreamEvents`; the diagnostic drain path never emits it. If you already consume
the raw `StreamEventsAsync` (or the client-level stream), the
`AsStreamItemsAsync()` extension projects any `IAsyncEnumerable<MxEvent>` into
the same `MxEventStreamItem` surface.
For alarms, the client exposes `QueryActiveAlarmsAsync` (one-shot snapshot of
the active alarms the gateway's central monitor currently holds),
`StreamAlarmsAsync` (server-streaming feed of alarm-state-change messages