Merge pull request 'A node says when it composes again, says what a long pass waits on, and composes on request (issue 774)' (#185) from fix/compose-retries-transient into main

cmd/mesh-host/engineseat.go

@@ -18,6 +18,7 @@ import (
 	"github.com/novox/mesh-host/internal/accounts"
 	"github.com/novox/mesh-host/internal/apply"
 	"github.com/novox/mesh-host/internal/link"
+	"github.com/novox/mesh-host/internal/selfcompose"
 	"github.com/novox/mesh-host/internal/store"
 )
 
@@ -292,3 +293,39 @@ func compositionFrom(current *agentNow) func(module, digest string, diff bool) (
 		return ag.Composition(ctx, module, digest, diff)
 	}
 }
+
+// composeReadFor bounds the verb compose's wait for its pass: a pass that waits on a seat longer says what it waits on.
+const composeReadFor = 2 * time.Minute
+
+// composeFrom answers the verb compose (issue 774): a pass of the composer that runs now, and with module that module's
+// apply tried again by the node agent; refused in words while the node composes nothing of its own.
+func composeFrom(current *agentNow, composer *atomic.Pointer[selfcompose.Composer]) func(module string) (any, error) {
+	return func(module string) (any, error) {
+		c := composer.Load()
+		if c == nil {
+			return nil, errors.New("this node composes none of its own modules now (it applies the controller's " +
+				"compositions, or its node agent is starting), so there is no pass to make")
+		}
+		out := map[string]any{}
+		if module != "" {
+			// Only a failed apply that is not held is tried again; what runs is never taken away (review round 1: the
+			// verb has no warrant, ADR 0563 §2, and a hold for a failed backup is the operator's to release, ADR 0568 §4).
+			ag := current.get()
+			if ag == nil {
+				return nil, errors.New("this engine's node agent has not started, so no apply is tried again")
+			}
+			if err := ag.RetryFailed(module); err != nil {
+				return nil, err
+			}
+			out["retried"] = module
+		}
+		ctx, cancel := context.WithTimeout(context.Background(), composeReadFor)
+		defer cancel()
+		pass, err := c.ComposeNow(ctx)
+		if err != nil {
+			return nil, err
+		}
+		out["pass"] = pass
+		return out, nil
+	}
+}

cmd/mesh-host/main.go

@@ -1142,6 +1142,7 @@ func runLink(ctx context.Context, opts options) error {
 	go reconcileActs(ctx, queue.Acts)
 	current := &agentNow{}
 	queue.Acts.Composition = compositionFrom(current)
+	queue.Acts.Compose = composeFrom(current, &composerNow)
 	setRunsHere(runsFromAgent(current))
 	factsOfNow := func(ctx context.Context, ag *agent.Agent) {
 		sayFacts(ctx, opts, mine.Membership.Signer, queue, say, ag)

cmd/mesh-host/nodeagent.go

@@ -668,6 +668,7 @@ func nodeAgent(ctx context.Context, opts options, signer ed25519.PublicKey, trus
 		}
 		if composer != nil {
 			machinesNow.Store(composer) // the machines a failing network part points at, by the placements (issue 774)
+			composerNow.Store(composer) // what the verb compose asks a pass of, this run's alone
 			composes.set(delivery.ComposesOwn)
 			env.Composed, env.SelfComposed, env.Trusted, env.Policy = composer, true, nil, composer.Policy
 			env.HoldsData = composer.HoldsData // what runs from the kept declaration says Data as its manifest does
@@ -692,6 +693,7 @@ func nodeAgent(ctx context.Context, opts options, signer ed25519.PublicKey, trus
 		stopRun()
 		cancel()
 		current.set(nil)
+		composerNow.Store(nil)
 		if ctx.Err() != nil {
 			return
 		}
@@ -814,6 +816,10 @@ func busyModules() []string {
 // (writeHealthReport); nil on a node that does not compose its own, whose parts then name no target.
 var machinesNow atomic.Pointer[selfcompose.Composer]
 
+// composerNow is the composer of the run that composes the node's own now, nil while the node applies the controller's
+// compositions or no run is going: what the verb compose asks a pass of (issue 774).
+var composerNow atomic.Pointer[selfcompose.Composer]
+
 // keptRetired says the kept declaration is read no more: every module the seat's placement puts on this node has a have of
 // its own (placedAllHave), so what runs is the agent's alone (ADR 0550 §2, design 47's L3 s9).
 var keptRetired atomic.Bool

go.mod

@@ -3,7 +3,7 @@ module github.com/novox/mesh-host
 go 1.26.0
 
 require (
-	git.novox.be/novox/mesh-sdk/go v0.1.14-0.20261030233444-8ed3b872e0ba
+	git.novox.be/novox/mesh-sdk/go v0.1.14-0.20261030235330-dfd5d8295e0a
 	github.com/nats-io/nats-server/v2 v2.11.17
 	github.com/nats-io/nats.go v1.54.0
 	golang.org/x/crypto v0.57.0

go.sum

@@ -1,5 +1,5 @@
-git.novox.be/novox/mesh-sdk/go v0.1.14-0.20261030233444-8ed3b872e0ba h1:vRjCeYiVsX8Q8FrT8IHa3z+zpzh2WVaDCnu7l/XYATI=
-git.novox.be/novox/mesh-sdk/go v0.1.14-0.20261030233444-8ed3b872e0ba/go.mod h1:ouERNpZizzwZ0HLEtAypsY9FR20FyUgUYRni6J7kVCY=
+git.novox.be/novox/mesh-sdk/go v0.1.14-0.20261030235330-dfd5d8295e0a h1:tjbRfUfE6pIylHFXg53SI2Sfgi1+QzWUyaY00xz25z4=
+git.novox.be/novox/mesh-sdk/go v0.1.14-0.20261030235330-dfd5d8295e0a/go.mod h1:ouERNpZizzwZ0HLEtAypsY9FR20FyUgUYRni6J7kVCY=
 github.com/antithesishq/antithesis-sdk-go v0.7.0-default-no-op h1:Z/MZK75wC/NSrkgqeNIa7jexam9uWzhLmFTSCPI/kn0=
 github.com/antithesishq/antithesis-sdk-go v0.7.0-default-no-op/go.mod h1:FQyySiasQQM8735Ddel3MRojmy4dA1IqCeyJ5jmPMbI=
 github.com/google/go-tpm v0.9.8 h1:slArAR9Ft+1ybZu0lBwpSmpwhRXaa85hWtMinMyRAWo=

internal/agent/agent.go

@@ -210,6 +210,8 @@ type Agent struct {
 	retry     chan []string
 	// reconcile carries the judge's edges: a resource found not active (ADR 0568 §2).
 	reconcile chan []string
+	// failedAgain carries the modules whose failed apply a person asked to try again (RetryFailed).
+	failedAgain chan string
 	// tries counts the transient failures in a row of each module; again carries a retry whose wait ran out, and
 	// transient asks for every transiently failed module at once. stopped is closed when Run returns.
 	tries     map[string]int
@@ -256,7 +258,7 @@ func New(env Env) *Agent {
 	}
 	return &Agent{env: env, wants: map[string]wantNow{}, comps: map[string]compNow{}, running: map[string]delivery.Have{}, applied: map[string]compNow{}, had: map[string]compNow{}, keptData: map[string]bool{},
 		attempt: map[string]attempt{}, checked: map[string]bool{}, prior: map[string]bool{}, said: map[string]string{}, notComposed: map[string]notComposedMark{},
-		retry: make(chan []string, 8), reconcile: make(chan []string, 64), tries: map[string]int{}, again: make(chan againAsk, 64),
+		retry: make(chan []string, 8), reconcile: make(chan []string, 64), failedAgain: make(chan string, 8), tries: map[string]int{}, again: make(chan againAsk, 64),
 		transient: make(chan struct{}, 1), stopped: make(chan struct{})}
 }
 
@@ -302,6 +304,15 @@ func (a *Agent) Run(ctx context.Context) error {
 				dirty[m] = true
 			}
 			a.mu.Unlock()
+		case m := <-a.failedAgain:
+			// A failed apply a person asked to try again: only one still failed and not held (RetryFailed); what runs is
+			// left as it is.
+			a.mu.Lock()
+			if att, ok := a.attempt[m]; ok && !att.Held {
+				delete(a.attempt, m)
+				dirty[m] = true
+			}
+			a.mu.Unlock()
 		case ask := <-a.again:
 			// The wait of a transiently failed apply ran out: tried again from the keys it has, if nothing tried it since.
 			a.mu.Lock()
@@ -1150,6 +1161,34 @@ func (a *Agent) Placed(except string) map[string]bool {
 	return out
 }
 
+// RetryFailed asks a module's failed apply to be tried again (the node-engine's verb compose, issue 774): only an apply
+// that failed and is not held for a failed backup, whose release is the operator's answer (ADR 0568 §4); what runs is
+// never taken away, so nothing is applied that the engine would not apply on its own at the next change. Refused in
+// words otherwise.
+func (a *Agent) RetryFailed(module string) error {
+	a.mu.Lock()
+	w := a.wants[module]
+	att, tried := a.attempt[module]
+	r, ran := a.running[module]
+	a.mu.Unlock()
+	switch {
+	case !w.Present || w.Want.Removed:
+		return fmt.Errorf("%s has no want on this machine, so there is no apply of it to try again", module)
+	case tried && att.Held:
+		return fmt.Errorf("%s is held for a failed backup: the operator's answer releases it", module)
+	case !tried && ran && r.FP == w.Want.FP && r.Phase != delivery.PhaseRemoved:
+		return fmt.Errorf("%s runs what its want names, nothing to retry", module)
+	case !tried:
+		return fmt.Errorf("%s has no failed apply to try again: it waits for its composition, or settles on its own", module)
+	}
+	select {
+	case a.failedAgain <- module:
+		return nil
+	default:
+		return fmt.Errorf("%s's retry is not taken now: the engine has too many asked; ask again in a moment", module)
+	}
+}
+
 // RemovedByWant is the modules whose want says removed (apply.Scope.RemovedByWant, issue 774): in the node's own apply,
 // only their resources may go. Called without a.mu held.
 func (a *Agent) RemovedByWant() map[string]bool {

internal/agent/agent_test.go

@@ -457,13 +457,27 @@ func TestAFailedApplyIsRetriedOnlyByAChangeOfEitherKey(t *testing.T) {
 		if n := len(b.appliedList()); n != 1 {
 			t.Fatalf("a failed apply was retried by an unrelated key or the clock (%d applies)", n)
 		}
-		// The want is written again: the retry.
+		// A person asks it again (the node-engine's verb compose, issue 774): the failed apply is tried again with no key
+		// changed; once it runs, nothing is left to retry, and a module with no want has none.
 		b.mu.Lock()
 		b.failWith = ""
 		b.mu.Unlock()
-		b.wants.put(wk, wraw)
-		eventually(t, "the change of the want did not retry", func() bool { return len(b.appliedList()) == 2 })
+		if err := b.ag.RetryFailed("web"); err != nil {
+			t.Fatalf("a failed apply was refused a retry: %v", err)
+		}
+		eventually(t, "the asked retry did not apply", func() bool { return len(b.appliedList()) == 2 })
 		eventually(t, "not running", func() bool { return b.haves.last("web").Phase == string(delivery.PhaseRunning) })
+		for module, words := range map[string]string{"web": "runs what its want names", "ghost": "has no want"} {
+			if err := b.ag.RetryFailed(module); err == nil || !strings.Contains(err.Error(), words) {
+				t.Errorf("a retry of %s: %v, want refused saying %q", module, err, words)
+			}
+		}
+		// The want is written again: no apply, it runs what the want names.
+		b.wants.put(wk, wraw)
+		time.Sleep(50 * time.Millisecond)
+		if n := len(b.appliedList()); n != 2 {
+			t.Errorf("a module running its want was applied again (%d applies)", n)
+		}
 	})
 	// The bus did not answer for a moment (mesh-vault on the bus node, 2026-10-25 23:04:24): tried again with no new
 	// revision of either key, by the backoff or by RetryTransient once the engine knows the bus is back; the have says
@@ -540,6 +554,17 @@ func TestADataModuleIsBackedUpFirstAndHeldWhenItFails(t *testing.T) {
 	if got := b.appliedList(); len(got) != 0 {
 		t.Fatalf("something was applied behind a failed backup: %v", got)
 	}
+	// A person's retry (the verb compose, issue 774) never releases it: the operator's answer does (ADR 0568 §4).
+	if err := b.ag.RetryFailed("db"); err == nil || !strings.Contains(err.Error(), "the operator's answer releases it") {
+		t.Errorf("a retry of db held for a failed backup: %v", err)
+	}
+	time.Sleep(50 * time.Millisecond)
+	b.mu.Lock()
+	asked := len(b.backups)
+	b.mu.Unlock()
+	if got := b.appliedList(); len(got) != 0 || asked != 1 {
+		t.Fatalf("a retry released db held for a failed backup: applied %v, %d backups asked", got, asked)
+	}
 
 	// The answer arrives in the module's own want: back up again (it fails again), then skip.
 	wk, again := wantFor(t, "db", fp2, dig2, delivery.BackupAgain)

internal/link/handacts.go

@@ -40,6 +40,8 @@ const (
 	ActSetuidSearch     = "setuid-search"
 	// ActComposition reads a module's composition beside what runs of it; read-only, no warrant (issue 774).
 	ActComposition = "composition"
+	// ActCompose composes the node now, the recovery of a failed composition (issue 774).
+	ActCompose = "compose"
 )
 
 // The states of an act.
@@ -132,6 +134,9 @@ type HandActs struct {
 	// Composition reads one module's composition beside what runs of it, read-only (the verb composition, issue 774):
 	// the latest of its want's build or the one whose digest is given, and with diff the resources that differ.
 	Composition func(module, digest string, diff bool) (any, error)
+	// Compose makes a pass of the node's composition now and says how it went; with module, its apply is also tried
+	// again (the verb compose, issue 774).
+	Compose func(module string) (any, error)
 	// Say says one of the seat's events, asked, done or refused; nil says nothing.
 	Say func(event string, body any)
 	Now func() time.Time
@@ -190,6 +195,11 @@ func (h *HandActs) Call(c EngineCall) (any, error) {
 			diff = b
 		}
 		return h.Composition(arg("module"), arg("digest"), diff)
+	case ActCompose:
+		if h.Compose == nil {
+			return nil, errors.New("this engine composes nothing on a call yet: its node agent has not started")
+		}
+		return h.Compose(arg("module"))
 	case "acts":
 		if id := arg("id"); id != "" {
 			return h.Act(id)

internal/link/handover_test.go

@@ -239,7 +239,8 @@ func TestALinkThatHearsHandOversIsGivenTheQueues(t *testing.T) {
 			JudgeDirectories: func(by string) (string, error) { return "judged for " + by, nil },
 			Composition: func(module, digest string, diff bool) (any, error) {
 				return map[string]any{"module": module, "digest": digest, "diff": diff}, nil
-			}}
+			},
+			Compose: func(module string) (any, error) { return map[string]any{"composed": true, "retried": module}, nil }}
 		for _, row := range []struct{ verb, subject, want string }{
 			{"judge-directories", "mesh.seat.node-engine.call.judge-directories.n1.shanks~node-tools", `"said":"judged for shanks.node-tools"`},
 			{"judge-directories", EngineToolSubject("judge-directories", "n1"), `"said":"judged for unproven"`},
@@ -248,6 +249,9 @@ func TestALinkThatHearsHandOversIsGivenTheQueues(t *testing.T) {
 			// The read verb composition (issue 774) takes its module, digest and diff, as a word or a boolean.
 			{"composition", EngineToolSubject("composition", "n1") + " " + `{"module":"postgres","digest":"sha256:a","diff":true}`, `"diff":true,"digest":"sha256:a","module":"postgres"`},
 			{"composition", EngineToolSubject("composition", "n1") + " " + `{"module":"postgres","diff":"true"}`, `"diff":true,"digest":"","module":"postgres"`},
+			// The recovery verb compose (issue 774) makes a pass now, with a module's apply tried again when named.
+			{"compose", EngineToolSubject("compose", "n1"), `"composed":true,"retried":""`},
+			{"compose", EngineToolSubject("compose", "n1") + " " + `{"module":"route-proxy"}`, `"retried":"route-proxy"`},
 			{"unserved", "mesh.seat.node-engine.call.unserved.n1.x", `"error":"the node seat node-engine serves no verb \"unserved\""`},
 		} {
 			subject, body, ok := strings.Cut(row.subject, " ")

internal/selfcompose/selfcompose.go

@@ -127,6 +127,72 @@ type Composer struct {
 	missed atomic.Uint64
 	// owed is set when a pass refused for a bus that did not answer: the composer composes again after a bound.
 	owed atomic.Bool
+	// tries counts the passes owed in a row, and wait is the bound before the next (LaterWait): set by the pass that owes
+	// one, so the refusal it marks says when the node composes again (issue 774, 2026-10-31). Under mu.
+	tries int
+	wait  time.Duration
+	// asked are the compose verb's asks (ComposeNow), each answered once the pass it asks is made.
+	asked chan chan Pass
+	// last is how the last pass went, under mu.
+	last Pass
+	// watching says the composer's loop runs (Watch), so an ask of a pass can be answered.
+	watching atomic.Bool
+}
+
+// Pass is how one pass went, as the node-engine's verb compose answers it.
+type Pass struct {
+	// At is when the pass ended.
+	At time.Time `json:"at"`
+	// Refused is why the node composed none of its modules; empty when it composed.
+	Refused string `json:"refused,omitempty"`
+	// Held is each module refused alone, with why.
+	Held map[string]string `json:"held,omitempty"`
+	// Handed is how many compositions the pass handed the agent that differ from what it handed before.
+	Handed int `json:"handed"`
+	// Again is when the node composes again on its own, after a seat or a read that did not answer; zero when it does
+	// not need to.
+	Again time.Time `json:"again,omitempty"`
+	// Earlier says the answer is the last pass, made less than askedWithin ago, and no new one was made for the ask.
+	Earlier bool `json:"earlier,omitempty"`
+}
+
+// askedWithin is how recent a pass answers an ask of the verb compose without a new one: the first wait after a seat
+// that did not answer (laterAgain's first).
+var askedWithin = 5 * time.Second
+
+// passSaysEvery is how often a pass that runs on says, in the journal, how long it has run and what it waits on.
+var passSaysEvery = time.Minute
+
+// owe marks a pass owed after a seat or a read that did not answer: the next wait of the growing bound (LaterWait), and
+// the words that say when the node composes again, for the refusal it marks.
+func (c *Composer) owe() string {
+	c.mu.Lock()
+	defer c.mu.Unlock()
+	c.wait = LaterWait(c.tries)
+	c.tries++
+	c.owed.Store(true)
+	return fmt.Sprintf("the node composes again at %s (try %d)", time.Now().Add(c.wait).Format("15:04:05"), c.tries)
+}
+
+// paid clears what owe marked: the pass composed, or owes nothing.
+func (c *Composer) paid() {
+	c.mu.Lock()
+	defer c.mu.Unlock()
+	c.tries, c.wait = 0, 0
+	c.owed.Store(false)
+}
+
+// againLine is what a retry pass says as it starts: the try its refusal named.
+func (c *Composer) againLine() string {
+	_, try := c.waitNow()
+	return fmt.Sprintf("the node composes again now (try %d), after a seat or a read that did not answer", try)
+}
+
+// waitNow is the bound before the pass owed, and its try.
+func (c *Composer) waitNow() (time.Duration, int) {
+	c.mu.Lock()
+	defer c.mu.Unlock()
+	return c.wait, c.tries
 }
 
 // note counts err when it is the bus not answering, and returns it.
@@ -150,7 +216,39 @@ func New(env Env) *Composer {
 	}
 	return &Composer{env: env, wants: map[string]delivery.Want{}, placements: map[string]delivery.Placement{},
 		records: map[string]delivery.NodeRecord{}, inOwn: map[string]bool{}, modules: map[string]delivery.Module{}, values: map[string]map[string]any{},
-		manifests: map[string][]byte{}, archives: map[string][]byte{}, handed: map[string]string{}, sealedOwn: map[string]ownSealed{}}
+		manifests: map[string][]byte{}, archives: map[string][]byte{}, handed: map[string]string{}, sealedOwn: map[string]ownSealed{},
+		asked: make(chan chan Pass, 8)}
+}
+
+// ComposeNow asks the composer for a pass now (the node-engine's verb compose, issue 774): the growing bound after a seat
+// that did not answer starts again, and the answer is how the pass went. Refused in words when no pass is made before
+// ctx ends: the composer not running, or its inputs not all read yet.
+//
+// A pass ended less than askedWithin ago is the answer, and no new one is made, so a caller asking in a loop never
+// drives passes back to back against the vault; an ask never restarts the growing bound either (review round 1).
+func (c *Composer) ComposeNow(ctx context.Context) (Pass, error) {
+	if !c.watching.Load() {
+		return Pass{}, errors.New("the node's composer runs no loop now: it composes nothing until its agent starts again")
+	}
+	c.mu.Lock()
+	last := c.last
+	c.mu.Unlock()
+	if !last.At.IsZero() && time.Since(last.At) < askedWithin {
+		last.Earlier = true
+		return last, nil
+	}
+	reply := make(chan Pass, 1)
+	select {
+	case c.asked <- reply:
+	case <-ctx.Done():
+		return Pass{}, errors.New("the node's composer took no ask: it composes nothing now")
+	}
+	select {
+	case p := <-reply:
+		return p, nil
+	case <-ctx.Done():
+		return Pass{}, fmt.Errorf("the node made no pass in time; it waits on %s", c.waitingOn())
+	}
 }
 
 // ownSealed is one own secret's sealed value at a generation and the key it is sealed to.
@@ -262,15 +360,17 @@ func (c *Composer) Watch(ctx context.Context, _ string) (<-chan agent.Update, er
 		}()
 	}
 	out := make(chan agent.Update, 64)
+	c.watching.Store(true)
 	go func() {
 		defer close(out)
+		defer c.watching.Store(false)
 		caught := make([]bool, len(ins))
 		ready := false
 		dirty := false
 		// A pass refused for a bus that did not answer is made again after a bound (laterAgain), not kept until another
 		// input moves (issue 1126, 2026-10-28).
 		var again <-chan time.Time
-		tries := 0
+		var replies []chan Pass
 		for {
 			// Drain whatever came, then compose once for all of it.
 			select {
@@ -279,6 +379,13 @@ func (c *Composer) Watch(ctx context.Context, _ string) (<-chan agent.Update, er
 			case <-again:
 				again = nil
 				dirty = true
+				c.env.Say(c.againLine())
+			case reply := <-c.asked:
+				// Asked by hand (the verb compose): composed now; the growing bound is not restarted, so a pass that
+				// fails again waits longer, not less.
+				replies = append(replies, reply)
+				again = nil
+				dirty = true
 			case t := <-merged:
 				if !t.ok {
 					c.env.Say(fmt.Sprintf("the watch of %s ended; the node composes no more until its agent starts again", ins[t.in].bucket))
@@ -318,13 +425,19 @@ func (c *Composer) Watch(ctx context.Context, _ string) (<-chan agent.Update, er
 				}
 			}
 			if c.owed.Load() {
-				wait := LaterWait(tries)
-				tries++
+				wait, _ := c.waitNow()
 				c.env.Say(fmt.Sprintf("the node composes again in %s: a read on the bus did not answer", wait))
 				again = time.After(wait)
 			} else {
-				tries, again = 0, nil
+				again = nil
 			}
+			c.mu.Lock()
+			last := c.last
+			c.mu.Unlock()
+			for _, r := range replies {
+				r <- last
+			}
+			replies = nil
 			if !c.caughtSent() {
 				select {
 				case out <- agent.Update{Caught: true}:
@@ -518,6 +631,35 @@ func (c *Composer) takeValue(u agent.Update) {
 // compose composes the node once and answers the updates to hand: each composition whose digest differs from the one
 // handed, and a deletion for each handed one no want names any more.
 func (c *Composer) compose(ctx context.Context) []agent.Update {
+	pass := Pass{}
+	defer func() {
+		pass.At = time.Now()
+		if c.owed.Load() {
+			wait, _ := c.waitNow()
+			pass.Again = pass.At.Add(wait)
+		}
+		c.mu.Lock()
+		c.last = pass
+		c.mu.Unlock()
+	}()
+	// A pass that runs on says so, and what it waits on, every passSaysEvery (issue 774, 2026-10-31: ace said nothing for
+	// five minutes while a pass waited after the vault restarted).
+	began := time.Now()
+	long, stopLong := context.WithCancel(ctx)
+	defer stopLong()
+	go func() {
+		t := time.NewTicker(passSaysEvery)
+		defer t.Stop()
+		for {
+			select {
+			case <-long.Done():
+				return
+			case <-t.C:
+				c.env.Say(fmt.Sprintf("the node's composition has run for %s and waits on %s",
+					time.Since(began).Round(time.Second), c.waitingOn()))
+			}
+		}
+	}()
 	c.mu.Lock()
 	fps := map[string]string{}
 	kept := map[string]string{}
@@ -542,7 +684,7 @@ func (c *Composer) compose(ctx context.Context) []agent.Update {
 		// The catalogue no longer places it: the agent returns to the controller's compositions (endWhenUnplaced), and
 		// this is no refusal of what it runs (issue 1126, the seed's revert).
 		c.sayOnce("this node's key in the seat's placement state is gone: it composes nothing of its own, and keeps running what it runs")
-		c.owed.Store(false) // nothing of its own to compose, so no pass is owed and no timer armed again
+		c.paid() // nothing of its own to compose, so no pass is owed and no timer armed again
 		return nil
 	}
 	// Composed from the build its want names (frombus.Scope.KeptAt, issue 1126).
@@ -581,11 +723,20 @@ func (c *Composer) compose(ctx context.Context) []agent.Update {
 		return nil
 	}
 	if err != nil {
+		why := fmt.Sprintf("%v", err)
+		// A bus that did not answer past every try is no answer: composed again after a bound (issue 1126, 2026-10-28),
+		// and the refusal says when (issue 774, 2026-10-31). A refusal is final until an input moves.
+		if BusFailed(err) {
+			why += "; " + c.owe()
+		} else {
+			c.paid()
+		}
+		pass.Refused = why
 		// Said before the agent marks the modules: the marks read each have first, which a paused bus delays.
-		c.sayOnce(fmt.Sprintf("the node cannot compose its modules, and keeps running what it runs: %v", err))
-		c.refused(fmt.Sprintf("%v", err), nil)
-		// A bus that did not answer past every try is no answer: composed again after a bound (issue 1126, 2026-10-28).
-		c.owed.Store(BusFailed(err))
+		c.sayOnce("the node cannot compose its modules, and keeps running what it runs: " + why)
+		done := c.waiting("the agent marking each wanted module's have")
+		c.refused(why, nil)
+		done()
 		return nil
 	}
 	// The boxes it runs, slot by slot: read before c.mu is taken, since the agent's lock is taken to read what runs.
@@ -596,14 +747,17 @@ func (c *Composer) compose(ctx context.Context) []agent.Update {
 	if c.missed.Load() > missed {
 		// A have of what runs the bus did not answer for: its boxes would be the vault's, sealed again, and the
 		// composition another; nothing is handed, and the pass is made again after a bound (issue 1126, 2026-10-28).
-		c.sayOnce("the node composes nothing new for now, and keeps running what it runs: the have of what it runs did not answer on the bus")
-		c.owed.Store(true)
+		pass.Refused = "the have of what it runs did not answer on the bus; " + c.owe()
+		c.sayOnce("the node composes nothing new for now, and keeps running what it runs: " + pass.Refused)
 		return nil
 	}
-	c.owed.Store(false)
+	c.paid()
+	pass.Held = held
 	// A module refused alone is held: it keeps what it runs, its have says why, and every other module is applied (issue
 	// 1126: one module's refusal froze the whole node).
+	done := c.waiting("the agent marking each held module's have")
 	c.refused("", held)
+	done()
 	if len(held) == 0 {
 		c.sayOnce("")
 	} else {
@@ -658,6 +812,7 @@ func (c *Composer) compose(ctx context.Context) []agent.Update {
 			out = append(out, agent.Update{Key: key, Deleted: true, Revision: c.rev})
 		}
 	}
+	pass.Handed = len(out)
 	return out
 }
 

internal/selfcompose/selfcompose_test.go

@@ -6,6 +6,7 @@ import (
 	"encoding/base64"
 	"encoding/hex"
 	"encoding/json"
+	"errors"
 	"fmt"
 	"os"
 	"slices"
@@ -238,6 +239,29 @@ func TestTheNodeComposesItsWantsFromTheBus(t *testing.T) {
 		}
 	}
 	b.mu.Unlock()
+	// Issue 774, 2026-10-31: the verb compose makes a pass now and says how it went; a composer that runs no watch
+	// makes none, and says so rather than waiting.
+	// The pass made at the start is less than askedWithin old for a moment: the ask waits that out first.
+	setFor(t, &askedWithin, 200*time.Millisecond)
+	time.Sleep(250 * time.Millisecond)
+	madeBefore := composed.Load()
+	cctx, ccancel := context.WithTimeout(ctx, 5*time.Second)
+	nowPass, err := c.ComposeNow(cctx)
+	ccancel()
+	if err != nil || nowPass.At.IsZero() || nowPass.Refused != "" || !nowPass.Again.IsZero() || composed.Load() == madeBefore {
+		t.Errorf("compose now: %+v (%v), %d pass(es) made; want one pass, composed", nowPass, err, composed.Load()-madeBefore)
+	}
+	// Asked again at once, the last pass is the answer: a caller in a loop drives no passes back to back (review round 1).
+	madeBefore = composed.Load()
+	if again, err := c.ComposeNow(ctx); err != nil || !again.Earlier || !again.At.Equal(nowPass.At) || composed.Load() != madeBefore {
+		t.Errorf("compose now asked again at once: %+v (%v), %d pass(es) made; want the last pass, none made", again, err, composed.Load()-madeBefore)
+	}
+	// A composer whose loop runs not refuses at once, not after the verb's whole bound.
+	idle, icancel := context.WithTimeout(ctx, 5*time.Second)
+	if _, err := New(env).ComposeNow(idle); err == nil || !strings.Contains(err.Error(), "runs no loop") || idle.Err() != nil {
+		t.Errorf("a composer that watches nothing answered compose now: %v", err)
+	}
+	icancel()
 	lemursKey, _ := delivery.ComposedKey("n1", "lemurs", fp)
 	nodeKey, _ := delivery.ComposedNodeKey("n1")
 	if !got[lemursKey] || !got[nodeKey] || len(got) != 2 {
@@ -277,15 +301,27 @@ func TestTheNodeComposesItsWantsFromTheBus(t *testing.T) {
 		c.mu.Unlock()
 		return c2, &refusals
 	}
+	setFor(t, &passSaysEvery, 5*time.Millisecond)
 	for name, row := range map[string]struct {
 		failures int // reads failing before the bus answers
 		cancel   bool
 		composes bool
+		fail     error         // what each failing read says; JetStream not enabled when nil
+		slow     time.Duration // each failing read takes this long
+		// again: the refusal says when the node composes again, and a pass is owed (issue 774, 2026-10-31: ace's
+		// route-proxy said only "the vault's own could not be asked ... context deadline exceeded").
+		again bool
 	}{
 		"the bus paused for one read":   {failures: 1, composes: true},
-		"the bus paused past every try": {failures: 100},
+		"the bus paused past every try": {failures: 100, again: true},
 		"the run ended mid-pass":        {failures: 100, cancel: true},
+		"the vault did not answer in time": {failures: 100, again: true, slow: 30 * time.Millisecond,
+			fail: fmt.Errorf("the vault's own could not be asked: no answer from mesh.seat.mesh-vault.call.own.node~n1: %w", context.DeadlineExceeded)},
+		"a refusal": {failures: 100, fail: errors.New("nats says ${dir:mesh-state}, and nats declares no directory")},
 	} {
+		b.mu.Lock()
+		saidBefore := len(b.said)
+		b.mu.Unlock()
 		failed := 0
 		pctx, pcancel := context.WithCancel(ctx)
 		c2, refusals := fresh(func(gctx context.Context, bucket, key string) ([]byte, error) {
@@ -295,6 +331,10 @@ func TestTheNodeComposesItsWantsFromTheBus(t *testing.T) {
 					pcancel()
 					return nil, fmt.Errorf("the vault's own could not be asked: %w", context.Canceled)
 				}
+				time.Sleep(row.slow)
+				if row.fail != nil {
+					return nil, fmt.Errorf("the revisions of %s in %s: %w", key, bucket, row.fail)
+				}
 				return nil, fmt.Errorf("the revisions of %s in %s: %w", key, bucket, nats.ErrJetStreamNotEnabled)
 			}
 			b.mu.Lock()
@@ -303,14 +343,39 @@ func TestTheNodeComposesItsWantsFromTheBus(t *testing.T) {
 		})
 		handed := c2.compose(pctx)
 		pcancel()
+		tries := 1 + len(composeAgain) // a bus paused for a moment is tried again at once; a timeout or a refusal is not
+		if row.fail != nil {
+			tries = 1
+		}
 		switch {
 		case row.composes && (len(handed) == 0 || len(*refusals) != 0):
 			t.Errorf("%s: handed %d, refusals %q; want composed and nothing refused", name, len(handed), *refusals)
-		case !row.composes && !row.cancel && (len(*refusals) != 1 || failed != 1+len(composeAgain)):
-			t.Errorf("%s: refusals %q after %d reads; want one refusal after %d tries", name, *refusals, failed, 1+len(composeAgain))
+		case !row.composes && !row.cancel && (len(*refusals) != 1 || failed != tries):
+			t.Errorf("%s: refusals %q after %d reads; want one refusal after %d tries", name, *refusals, failed, tries)
 		case row.cancel && len(*refusals) != 0:
 			t.Errorf("%s: a pass whose run ended refused %q", name, *refusals)
 		}
+		if !row.composes && !row.cancel && len(*refusals) > 0 {
+			// The retry pass says the same try the refusal named (review round 1: it said try 2 for try 1).
+			says := strings.Contains((*refusals)[0], "; the node composes again at ") && strings.HasSuffix((*refusals)[0], "(try 1)") &&
+				strings.Contains(c2.againLine(), "(try 1)")
+			if says != row.again || c2.owed.Load() != row.again || (c2.last.Again.IsZero() == row.again) || c2.last.Refused != (*refusals)[0] {
+				t.Errorf("%s: refused %q, owes a pass %v, last pass %+v; want the time it composes again said and owed: %v",
+					name, (*refusals)[0], c2.owed.Load(), c2.last, row.again)
+			}
+		}
+		// A pass that runs on says so, and what it waits on (issue 774, 2026-10-31: ace was silent for five minutes).
+		if row.slow > 0 {
+			b.mu.Lock()
+			long := false
+			for _, line := range b.said[saidBefore:] {
+				long = long || strings.Contains(line, "the node's composition has run for ") && strings.Contains(line, " and waits on ")
+			}
+			b.mu.Unlock()
+			if !long {
+				t.Errorf("%s: a pass that ran on said nothing of what it waits on", name)
+			}
+		}
 	}
 	if !Transient(fmt.Errorf("the manifest of avahi: the revisions of x in mesh-delivery-indexer_manifests: nats: jetstream not enabled")) ||
 		!Transient(fmt.Errorf("x: %w", nats.ErrNoResponders)) || Transient(fmt.Errorf("nats says ${dir:mesh-state}, and nats declares no directory")) ||
@@ -881,7 +946,8 @@ func TestTheNodeComposesItsWantsFromTheBus(t *testing.T) {
 	// Review of L2b, round 1, finding 2: the candidates' state is not there when the node starts (a delivery before the
 	// indexer made it): the shadow tries again, and answers once it appears.
 	sb.missing(delivery.PlacementCandidatesBucket, true)
-	go sh.Keep(sctx, func(int) time.Duration { return 5 * time.Millisecond })
+	kept := make(chan struct{})
+	go func() { sh.Keep(sctx, func(int) time.Duration { return 5 * time.Millisecond }); close(kept) }()
 	time.Sleep(30 * time.Millisecond)
 	sb.missing(delivery.PlacementCandidatesBucket, false)
 	sb.made(7)
@@ -1216,6 +1282,9 @@ func TestTheNodeComposesItsWantsFromTheBus(t *testing.T) {
 		t.Errorf("a shadow not ready said %q", nsaid)
 	}
 	nmu.Unlock()
+	// The first shadow ends before slowAfter is set: it reads it while it answers (a race the race detector names).
+	scancel()
+	<-kept
 	setFor(t, &slowAfter, time.Nanosecond) // every answer is slow: said, and noted in words that do not move with the time
 	nb.send(delivery.AssignerWantsBucket, agent.Update{Caught: true})
 	select {