perf: persist remote runtime events in one transaction per checkpoint

Remote semantic events were written with one or two synchronous SQLite transactions each before being forwarded; recoverable conversations also rewrote the whole assistant message per event. Events now commit together at each remote checkpoint (before the Agent can acknowledge), and before non-remote, question and terminal events, on a size cap or a 50 ms timer. Only newly stored events are forwarded, in order; a failed batch rolls back and blocks later commits. Applies to interactive, background and recovery runs.

src/main/remote-event-batcher.test.ts

@@ -0,0 +1,66 @@
+import { afterEach, describe, expect, it, vi } from 'vitest'
+import { RemoteEventBatcher } from './remote-event-batcher'
+
+const provenance = (eventIndex: number) => ({
+  source: 'remote-semantic-transcript' as const,
+  bindingId: 'binding',
+  operationId: 'operation',
+  semanticSequence: '1',
+  eventIndex
+})
+
+describe('RemoteEventBatcher', () => {
+  afterEach(() => {
+    vi.useRealTimers()
+  })
+
+  it('commits buffered events once and reports them in order', () => {
+    const persist = vi.fn((entries: readonly { event: string }[]) =>
+      entries.map(({ event }) => event !== 'replayed'))
+    const committed: Array<[string, boolean]> = []
+    const batcher = new RemoteEventBatcher<string>({ persist })
+    for (const [index, event] of ['a', 'replayed', 'b'].entries()) {
+      batcher.add(provenance(index), event, (value, inserted) => committed.push([value, inserted]))
+    }
+    expect(persist).not.toHaveBeenCalled()
+    batcher.flush()
+    batcher.flush()
+    expect(persist).toHaveBeenCalledOnce()
+    expect(committed).toEqual([['a', true], ['replayed', false], ['b', true]])
+    batcher.dispose()
+  })
+
+  it('flushes at the size cap and on the safety timer', () => {
+    vi.useFakeTimers()
+    const persist = vi.fn((entries: readonly unknown[]) => entries.map(() => true))
+    const batcher = new RemoteEventBatcher<number>({ persist, maximumEvents: 2, flushIntervalMs: 50 })
+    batcher.add(provenance(0), 0)
+    batcher.add(provenance(1), 1)
+    expect(persist).toHaveBeenCalledTimes(1)
+    batcher.add(provenance(2), 2)
+    vi.advanceTimersByTime(49)
+    expect(persist).toHaveBeenCalledTimes(1)
+    vi.advanceTimersByTime(1)
+    expect(persist).toHaveBeenCalledTimes(2)
+    expect(persist.mock.calls[1]![0]).toHaveLength(1)
+    batcher.dispose()
+  })
+
+  it('rethrows a failed commit and stays failed so later checkpoints are never committed', () => {
+    vi.useFakeTimers()
+    const onError = vi.fn()
+    const persist = vi.fn((): boolean[] => {
+      throw new Error('disk full')
+    })
+    const onCommitted = vi.fn()
+    const batcher = new RemoteEventBatcher<string>({ persist, onError })
+    batcher.add(provenance(0), 'a', onCommitted)
+    vi.advanceTimersByTime(50)
+    expect(onError).toHaveBeenCalledWith(new Error('disk full'))
+    expect(() => batcher.add(provenance(1), 'checkpoint')).toThrow('disk full')
+    expect(() => batcher.flush()).toThrow('disk full')
+    expect(persist).toHaveBeenCalledOnce()
+    expect(onCommitted).not.toHaveBeenCalled()
+    batcher.dispose()
+  })
+})