feat(agent): add queue-aware Agent.cancel() primitive

abort() only kills the in-flight step, so a queued-but-not-yet-started prompt
ran to completion after a cancel and a prompt accepted right after could be
batched into the cancelled turn (the loop merges queued messages into one turn).
This closes TODO(rfc010-cancel-prestep) with a distinct cancel() verb.

cancel() clears the queued + steering FIFOs, aborts the in-flight step, and
drives a turn-scoped marker on the LoopHandle that the driver checks at EVERY
point a turn could start or continue:
- right after the idle wait (window 1): drop the about-to-run turn and settle
  whenIdle() waiters directly (no running→idle transition fires, and no
  agent/status is emitted, so an ACP listener can't see a spurious idle that
  resolves a freshly-queued prompt as cancelled);
- after the synchronous setStatus('running') emit (window 2): a running listener
  can cancel in the gap before runTurn;
- in the step-start window (before runStep, after setAbort): a synchronous
  turn-start/step-start listener can cancel before any AbortController exists;
- at the continuation gate: a cancel during the continuation waterfall (the
  finished step's controller already cleared) ends the turn aborted.

The marker is ARMED only when there is something to cancel (running, an
in-flight step, or queued/steering work) — an idle no-op cancel cannot leave it
set to drop a later prompt — and RESET unconditionally once per loop iteration,
so it governs exactly one turn and never leaks onto the next prompt (even when a
send() lands in the cancelled turn's flush window).

ACP session/cancel now maps to agent.cancel() (keeping the synchronous
settlePrompt). Teardown/disconnect still use abort('disposed') until PR D, so
the ACP README narrows the remaining best-effort window to teardown only.

Tests (agent-loop/cancel.spec.ts) cover every window unit-level (the F1 hang
guard: a whenIdle() waiter registered before a pre-step cancel resolves; the F2
leak guard: idle cancel then a prompt runs; mid-step, continuation, both
pre-step windows, turn-start-listener, steering-cleared, marker-reset). ACP
turns.spec.ts adds the through-bridge tests with NO intervening whenIdle (idle
cancel→prompt runs; mid-stream cancel→immediate next prompt runs) and updates
the stale pre-step test to the queue-aware guarantee. The existing cancel
snapshot golden is byte-identical (it drives the new cancel() path end-to-end
through the real subprocess), so no new golden is needed. 100% coverage.
This commit is contained in:
Tianyi Cui
2026-06-20 04:51:32 +08:00
parent 3d67a98291
commit c4bc6e0e38
12 changed files with 504 additions and 31 deletions

View File

@@ -66,6 +66,8 @@ forever:
Error containment: a throwing plugin ends the **turn**, never the loop. Dispose mid-turn emits `agent/status('disposed')` and ends with reason `disposed`. A step that hits the model's output-token ceiling makes the turn end `max-tokens` (the rule: any `max-tokens` step in the turn surfaces as `max-tokens`; `disposed`/`aborted`/`error` still take precedence) — distinct from a clean `completed` stop.
Cancellation: `agent.abort()` aborts only the in-flight step; `agent.cancel()` is the broad verb — it clears the queued + steering FIFOs, aborts the in-flight step, and drives a turn-scoped marker the driver checks at every point a turn could start or continue (right after the idle wait, after the `running` flip, before each step, and at the continuation gate) so a turn about to start is dropped. A cancelled turn ends `aborted`; a queued-but-not-started prompt never runs and cannot be batched into the cancelled turn. The marker is reset once per loop iteration, so a cancel governs exactly one turn and never leaks onto a later prompt.
### What is NOT here
Everything that goes beyond "call the model, run the tools, repeat" belongs to plugins listening on the event taxonomy:

View File

@@ -26,6 +26,14 @@ export class ReactLoopAgent implements Agent {
private _status: AgentStatus = 'idle'
private currentAbort: AbortController | undefined
/**
* Turn-scoped cancel marker, set by {@link cancel} and read/cleared by the
* driver loop (via the LoopHandle) at every point a turn could start or
* continue. Armed ONLY when there is something to cancel (a running turn, an
* in-flight step, or queued/steering work), so an idle no-op cancel cannot
* leave it set to wrongly drop a later prompt.
*/
private cancelRequested = false
private disposed: Promise<void>
private resolveDisposed!: () => void
/** Resolves when the driver loop has fully exited (tests/disposal). */
@@ -176,6 +184,30 @@ export class ReactLoopAgent implements Agent {
this.currentAbort?.abort(reason ?? 'aborted')
}
cancel(reason?: string): void {
// Arm-gate: only mark a cancellation when there is actually work to cancel —
// a running turn, an in-flight step, or queued/steering work. An idle cancel
// with nothing pending is a true no-op; arming the marker then would wrongly
// drop the NEXT legitimate prompt (the marker is consumed only at the loop's
// turn-decision points, which an idle parked loop does not reach until woken
// by a real send()). Note the gate canNOT be `status === 'running'` alone:
// the pre-step window (a send() queued but the loop not yet flipped to
// running) has status `idle` with `hasQueued` true, and the marker exists
// precisely to cover it.
if (this._status === 'running' || this.currentAbort !== undefined || this.inbox.hasQueued || this.inbox.hasSteering) {
this.cancelRequested = true
}
// Drop all pending queued + steering work (un-started prompts never run; the
// cancelled turn's steering is not re-enqueued). Cleared directly even when
// the loop is parked in waitForQueued — there is no turn to stop and nothing
// left for the parked loop to run, so no wake is needed.
this.inbox.clear()
// Interrupt an in-flight step immediately (the running turn observes the
// abort and ends `aborted`). The marker covers the windows where no step is
// running (pre-step, continuation).
this.currentAbort?.abort(reason ?? 'cancelled')
}
/**
* Resolve once the agent has reached quiescence after settling out of
* `running`. If it is already disposed, awaits {@link done} (the loop-exit
@@ -218,6 +250,15 @@ export class ReactLoopAgent implements Agent {
setAbort: controller => void (this.currentAbort = controller),
disposed: this.disposed,
isDisposed: () => this._status === 'disposed',
isCancelled: () => this.cancelRequested,
clearCancel: () => { this.cancelRequested = false },
// Settle whenIdle() waiters WITHOUT a status transition — the pre-step
// cancel-skip path drops the about-to-run turn and re-parks without ever
// flipping running→idle, so a waiter registered in the pre-step window
// (status idle, hasQueued was true) would otherwise hang. This emits no
// agent/status, so an ACP agent/status listener never sees a spurious idle
// that would resolve a freshly-queued prompt as cancelled.
settleIdle: () => { this.settleIdleWaiters() },
})
// The disposer must be infallible: it runs inside the fiber's LIFO
// disposal chain, where a throw would skip later disposers (e.g. the

View File

@@ -52,6 +52,16 @@ export class Inbox {
return this.steeringMessages.splice(0)
}
/**
* Discard all pending messages (queued + steering) without delivering them —
* used by `cancel()`, which drops un-started work rather than draining it into
* a turn. Unlike `drainQueued`/`drainSteering`, the messages are thrown away.
*/
clear(): void {
this.queuedMessages.length = 0
this.steeringMessages.length = 0
}
/** Wait until a queued message arrives or `cancel` resolves. */
waitForQueued(cancel: Promise<void>): Promise<void> {
if (this.hasQueued) return Promise.resolve()

View File

@@ -107,6 +107,26 @@ export interface LoopHandle {
/** Resolves when the agent is disposed — unblocks the idle wait. */
disposed: Promise<void>
isDisposed(): boolean
/**
* Whether a `cancel()` is pending for the current turn. The driver checks this
* at every decision point where a turn could start or continue (right after
* the idle wait, after the `running` flip, before each step, and at the
* continuation gate) and drops the about-to-run / continuing turn. Reset once
* per loop iteration via {@link clearCancel} after the turn returns, so the
* marker governs exactly one cancellation and never leaks to a later prompt.
*/
isCancelled(): boolean
/** Clear the cancel marker (called once per iteration after the turn returns). */
clearCancel(): void
/**
* Settle pending `whenIdle()` waiters WITHOUT a status transition. Used by the
* pre-step cancel-skip path: it drops the about-to-run turn and re-parks at the
* idle wait, so no `running→idle` transition fires to settle a `whenIdle()`
* waiter that was registered in the pre-step window — this settles it directly
* (it emits no `agent/status`, so an ACP `agent/status` listener never sees a
* spurious idle that would resolve a freshly-queued prompt as cancelled).
*/
settleIdle(): void
}
/**
@@ -148,7 +168,33 @@ export async function runLoop(ctx: Context, agent: ReactLoopAgent, handle: LoopH
await agent.inbox.waitForQueued(handle.disposed)
if (handle.isDisposed()) break
// Pre-step cancel (window 1): a `cancel()` landed after a `send()` woke the
// idle wait but before we flip to `running`. Drop the about-to-run turn: the
// queued/steering work is already cleared by `cancel()`, and we settle any
// `whenIdle()` waiter DIRECTLY (no status transition fires here, so the
// running→idle settle never runs) WITHOUT emitting `agent/status` (an ACP
// listener must not see a spurious idle that resolves a freshly-queued prompt
// as cancelled). Clear the marker and re-park.
if (handle.isCancelled()) {
handle.clearCancel()
handle.settleIdle()
continue
}
handle.setStatus('running')
// Pre-step cancel (window 2): `setStatus('running')` emits `agent/status`
// SYNCHRONOUSLY, so a `running` listener can `cancel()` in the gap between the
// check above and `runTurn`. `cancel()` already cleared the queued FIFO, so
// drop the turn before it starts (runTurn would otherwise throw on an empty
// queue) and transition back to idle — `running` was already emitted, so a
// real `idle` transition (which also settles waiters) balances the status.
if (handle.isCancelled()) {
handle.clearCancel()
handle.setStatus('idle')
continue
}
// Re-derive the turn number from the log each iteration (do NOT keep a local
// counter): an idle `agent.inject()` can append its own one-shot turn while
// the loop waits above, so the next real turn must continue from whatever
@@ -170,8 +216,18 @@ export async function runLoop(ctx: Context, agent: ReactLoopAgent, handle: LoopH
} catch { /* contained: a throwing agent/error listener must not kill the driver */ }
}
// Reset the cancel marker UNCONDITIONALLY here, after the turn returns and
// before the next iteration's idle wait. NOT gated on the idle transition
// below: a `send()` that lands during the cancelled turn's flush window makes
// `hasQueued` true at the `setStatus('idle')` guard, so an idle-gated reset
// would never fire and the stale marker would wrongly drop that next prompt's
// turn. Resetting per iteration scopes the marker to exactly the turn that was
// cancelled.
handle.clearCancel()
// Steering that arrived too late to join this turn (turn-end listeners,
// flush) becomes a queued message — it must never be stranded.
// flush) becomes a queued message — it must never be stranded. (A cancelled
// turn already cleared its steering, so there is nothing to re-enqueue.)
for (const message of agent.inbox.drainSteering()) {
agent.inbox.enqueue(message)
}
@@ -323,6 +379,20 @@ async function runTurn(ctx: Context, agent: ReactLoopAgent, handle: LoopHandle,
const abort = new AbortController()
handle.setAbort(abort)
// Cancel landing in the step-start window: a synchronous `agent/turn-start`
// or `agent/step-start` listener (both fire before this point) can have
// called `cancel()`, and `runStep` would otherwise run a full extra step
// with no AbortController having observed it. Check the marker AFTER
// setAbort (so the next-iteration drain sees a clean controller) and before
// `runStep`: drop the step, end the turn `aborted`. closeStep balances the
// already-appended step/start.
if (handle.isCancelled()) {
handle.setAbort(undefined)
reason = { kind: 'aborted', reason: 'cancelled' }
closeStep()
break
}
let stepOutcome: { hadToolCalls: boolean; finish: FinishReason } | { error: Error }
try {
stepOutcome = await runStep(ctx, agent, turn, step, abort.signal)
@@ -382,6 +452,16 @@ async function runTurn(ctx: Context, agent: ReactLoopAgent, handle: LoopHandle,
// next iteration's drain records it.
if (!shouldContinue && agent.inbox.hasSteering) shouldContinue = true
// A cancel that landed during the continuation window — after the step's
// AbortController was cleared (setAbort(undefined)) but before the next
// step starts — has no controller to observe it, so the turn-scoped marker
// ends the turn here. cancel() also cleared the steering FIFO, so the
// override above did not re-arm continuation.
if (handle.isCancelled()) {
reason = { kind: 'aborted', reason: 'cancelled' }
break
}
if (!shouldContinue || handle.isDisposed()) {
/* v8 ignore next -- disposal during continuation-decision window is a narrow race; error-path disposal is covered elsewhere */
if (handle.isDisposed()) reason = { kind: 'disposed' }

View File

@@ -0,0 +1,279 @@
/**
* Tests for the queue-aware `Agent.cancel()` primitive (PR C). `cancel()` is the
* broad verb — it clears queued + steering work, aborts an in-flight step, and
* drops a turn about to start — whereas `abort()` kills only the current step.
* These tests exercise every window where a cancel can land (idle, pre-step,
* mid-step, continuation) and the marker's arm/reset rules that keep a cancel
* from leaking to a later prompt or hanging `whenIdle()`.
*
* @module dsh-agent-loop/tests/cancel
*/
import { describe, expect, it } from 'vitest'
import { Context } from 'cordis'
import LlmService from '@deepseek-ai/dsh-llm'
import SessionStore, { TurnEndReason } from '@deepseek-ai/dsh-session'
import SystemPrompt from '@deepseek-ai/dsh-system-prompt'
import ToolRegistry from '@deepseek-ai/dsh-tools'
import AgentRegistry from '@deepseek-ai/dsh-agent'
import AgentLoop, { ReactLoopAgent } from '@deepseek-ai/dsh-agent-loop'
import { MockAdapter, textResponse } from './mock-adapter.ts'
async function harness(adapter: MockAdapter) {
const ctx = new Context()
await ctx.plugin(LlmService)
await ctx.plugin(SessionStore)
await ctx.plugin(SystemPrompt)
await ctx.plugin(ToolRegistry)
await ctx.plugin(AgentRegistry)
await ctx.plugin(AgentLoop, { agents: [] })
ctx.llm.registerAdapter(['mock'], adapter)
return ctx
}
function send(agent: ReactLoopAgent, text: string) {
agent.send([{ type: 'text', text }])
}
/** Resolve on the agent's next idle transition (event-based, not status poll). */
function waitForIdle(ctx: Context, agent: ReactLoopAgent): Promise<void> {
return new Promise((resolve) => {
const dispose = ctx.on('agent/status', (subject, status) => {
if (subject === agent && status === 'idle') { dispose(); resolve() }
})
})
}
/** All user-message texts recorded in the log (to assert what actually ran). */
function userTexts(agent: ReactLoopAgent): string[] {
return agent.session.events
.filter(e => e.type === 'user/message')
.flatMap(e => e.type === 'user/message' ? e.data.content : [])
.flatMap(b => b.type === 'text' ? [b.text] : [])
}
describe('Agent.cancel()', () => {
it('cancel() on an idle agent with nothing queued is a no-op; the next prompt runs (F2 leak guard)', async () => {
const adapter = new MockAdapter([textResponse('reply')])
const ctx = await harness(adapter)
const agent = ctx.agentLoop.create('a1', { model: 'mock' })
// The loop is parked at the idle wait with nothing queued. A cancel here must
// NOT arm the marker — otherwise the next legitimate prompt would be dropped.
agent.cancel('nothing to cancel')
send(agent, 'real prompt')
await waitForIdle(ctx, agent)
// The prompt ran: its user message is in the log and one turn completed.
expect(userTexts(agent)).toEqual(['real prompt'])
expect(agent.session.events.some(e => e.type === 'turn/end')).toBe(true)
})
it('pre-step cancel drops the about-to-start turn (no turn is opened)', async () => {
const adapter = new MockAdapter([textResponse('should not run')])
const ctx = await harness(adapter)
const agent = ctx.agentLoop.create('a1', { model: 'mock' })
// send() queues synchronously (status still idle, loop microtask not yet
// resumed). Cancel in that pre-step window: the queued turn must not run.
send(agent, 'drop me')
agent.cancel('pre-step')
// Give the loop a chance to wake and process the cancel.
await new Promise(r => setTimeout(r, 30))
// No turn was opened — the queued prompt was dropped, never recorded.
expect(userTexts(agent)).toEqual([])
expect(agent.session.events.some(e => e.type === 'turn/start')).toBe(false)
expect(agent.status).toBe('idle')
})
it('a whenIdle() waiter registered BEFORE a pre-step cancel resolves (F1 hang guard)', async () => {
const adapter = new MockAdapter([textResponse('x')])
const ctx = await harness(adapter)
const agent = ctx.agentLoop.create('a1', { model: 'mock' })
// Queue work, then register a whenIdle() waiter while in the pre-step window
// (status idle, hasQueued true) — it does NOT take the fast path. Then cancel.
// The skip path must settle this waiter directly (no running→idle transition
// ever fires), or it would hang forever.
send(agent, 'q')
const idle = agent.whenIdle()
agent.cancel('pre-step')
// Must resolve (not hang). A timeout makes the failure a clear test failure.
await Promise.race([
idle,
new Promise((_r, reject) => setTimeout(() => { reject(new Error('whenIdle hung after pre-step cancel')) }, 1000)),
])
expect(agent.status).toBe('idle')
})
it('cancel() mid-step aborts the in-flight model call; the turn ends aborted', async () => {
const adapter = new MockAdapter(['hang'])
const ctx = await harness(adapter)
const agent = ctx.agentLoop.create('a1', { model: 'mock' })
const reasons: TurnEndReason[] = []
ctx.on('agent/turn-end', (_a, _t, reason) => void reasons.push(reason))
send(agent, 'go')
await new Promise(r => setTimeout(r, 30))
expect(agent.status).toBe('running')
agent.cancel('mid-step')
await waitForIdle(ctx, agent)
expect(reasons).toEqual([{ kind: 'aborted', reason: 'mid-step' }])
})
it('cancel() with no reason defaults to "cancelled" when aborting an in-flight step', async () => {
const adapter = new MockAdapter(['hang'])
const ctx = await harness(adapter)
const agent = ctx.agentLoop.create('a1', { model: 'mock' })
const reasons: TurnEndReason[] = []
ctx.on('agent/turn-end', (_a, _t, reason) => void reasons.push(reason))
send(agent, 'go')
await new Promise(r => setTimeout(r, 30))
agent.cancel() // no reason → default 'cancelled'
await waitForIdle(ctx, agent)
expect(reasons).toEqual([{ kind: 'aborted', reason: 'cancelled' }])
})
it('a prompt sent AFTER a cancelled turn settles runs normally (marker reset)', async () => {
const adapter = new MockAdapter(['hang', textResponse('second reply')])
const ctx = await harness(adapter)
const agent = ctx.agentLoop.create('a1', { model: 'mock' })
// First turn hangs; cancel it mid-step.
send(agent, 'first')
await new Promise(r => setTimeout(r, 30))
agent.cancel('cancel first')
await waitForIdle(ctx, agent)
// The marker must have been reset after the cancelled turn — a fresh prompt
// runs to completion rather than being dropped by a stale marker.
send(agent, 'second')
await waitForIdle(ctx, agent)
expect(userTexts(agent)).toContain('second')
// The second turn completed (its reply was streamed).
const reasons = agent.session.events.filter(e => e.type === 'turn/end')
expect(reasons.length).toBe(2)
})
it('cancel from a synchronous agent/turn-start listener drops the step (step-start window)', async () => {
const adapter = new MockAdapter([textResponse('should not stream')])
const ctx = await harness(adapter)
const agent = ctx.agentLoop.create('a1', { model: 'mock' })
// A turn-start listener fires BEFORE any AbortController is installed for the
// step. Cancelling there must still drop the step (the turn-scoped marker,
// not abort(), is what catches this) — no model step runs.
let streamed = false
ctx.on('agent/stream-chunk', () => { streamed = true })
const dispose = ctx.on('agent/turn-start', (subject) => {
if (subject === agent) agent.cancel('from turn-start')
})
const reasons: TurnEndReason[] = []
ctx.on('agent/turn-end', (_a, _t, reason) => void reasons.push(reason))
send(agent, 'go')
await waitForIdle(ctx, agent)
dispose()
// No step streamed (the model never ran), and the turn ended aborted.
expect(streamed).toBe(false)
expect(reasons).toEqual([{ kind: 'aborted', reason: 'cancelled' }])
})
it('cancel during the continuation window ends the turn aborted and runs no further step', async () => {
// A continuation-waterfall listener cancels DURING the continuation decision
// (the finished step's AbortController is already cleared), and votes to
// continue — but the turn-scoped marker checked right after must end the turn
// `aborted` and run NO second step.
const adapter = new MockAdapter([textResponse('one'), textResponse('two')])
const ctx = await harness(adapter)
const agent = ctx.agentLoop.create('a1', { model: 'mock' })
let steps = 0
ctx.on('agent/step-start', () => { steps += 1 })
const reasons: TurnEndReason[] = []
ctx.on('agent/turn-end', (_a, _t, reason) => void reasons.push(reason))
let continued = false
ctx.on('agent/turn-continuation', async (subject, _turn, _default, next) => {
if (subject === agent && !continued) {
continued = true
agent.cancel('from continuation')
return true // vote to continue — the post-waterfall marker check must override
}
return next()
})
send(agent, 'go')
await waitForIdle(ctx, agent)
// Only ONE step ran (the second was cancelled in the continuation window),
// and the turn ended aborted.
expect(steps).toBe(1)
expect(reasons).toEqual([{ kind: 'aborted', reason: 'cancelled' }])
})
it('cancel from a synchronous agent/status(running) listener drops the turn (window 2)', async () => {
const adapter = new MockAdapter([textResponse('should not run')])
const ctx = await harness(adapter)
const agent = ctx.agentLoop.create('a1', { model: 'mock' })
// setStatus('running') emits agent/status SYNCHRONOUSLY, so a running
// listener can cancel in the gap between the loop's pre-step check and
// runTurn. The second check (after the running flip) must drop the turn —
// runTurn would otherwise throw on the now-empty queue.
let streamed = false
ctx.on('agent/stream-chunk', () => { streamed = true })
const dispose = ctx.on('agent/status', (subject, status) => {
if (subject === agent && status === 'running') agent.cancel('from running listener')
})
send(agent, 'go')
await waitForIdle(ctx, agent)
dispose()
// No turn opened, no step streamed, and a later prompt still runs (the marker
// was reset).
expect(streamed).toBe(false)
expect(agent.session.events.some(e => e.type === 'turn/start')).toBe(false)
})
it("cancel clears the turn's steering — it is not re-enqueued as a fresh turn", async () => {
const adapter = new MockAdapter(['hang'])
const ctx = await harness(adapter)
const agent = ctx.agentLoop.create('a1', { model: 'mock' })
send(agent, 'go')
await new Promise(r => setTimeout(r, 30))
expect(agent.status).toBe('running')
// Steer (joins the running turn's steering FIFO), then cancel: the steering
// must be dropped, NOT re-enqueued as a new queued turn.
agent.steer([{ type: 'text', text: 'steer text' }])
agent.cancel('cancel with steering')
await waitForIdle(ctx, agent)
// After the cancelled turn settles, the agent is idle with NO follow-up turn
// started from the dropped steering.
await new Promise(r => setTimeout(r, 30))
expect(agent.status).toBe('idle')
const turnStarts = agent.session.events.filter(e => e.type === 'turn/start')
expect(turnStarts.length).toBe(1) // only the original (cancelled) turn
// The steering text was dropped — it never reached the log.
const flat = agent.session.events
.filter(e => e.type === 'steering/message')
.flatMap(e => e.type === 'steering/message' ? e.data.content : [])
.flatMap(b => b.type === 'text' ? [b.text] : [])
expect(flat).not.toContain('steer text')
})
})