Compare commits

..
Author SHA1 Message Date
Miguel Palhas 74a423567f docs(maestro): fix disabled-unstable-variant claim in STRUCTURED-VIEW.md
A SessionUpdate variant gated behind an unstable ACP feature this build
doesn't enable is compiled out of the enum, not decoded into a Notice.
It fails deserialization the same as a genuinely unknown future tag.
2026-09-05 05:35:19 +01:00
Miguel Palhas b225fbdd69 docs(maestro): narrow the Notice forward-compat claim in STRUCTURED-VIEW.md
Notice only covers SessionUpdate variants the agent-client-protocol
dependency already decodes but the projection doesn't special-case.
SessionUpdate is an internally-tagged enum with no unknown-variant
fallback, so a genuinely new wire variant fails deserialization before
the projection ever sees it, rather than becoming a Notice. A new
variant needs the dependency upgraded first.

Adds a projection test asserting that an unrecognised sessionUpdate
tag is a deserialization error.

Refs #186
2026-09-05 05:28:11 +01:00
Miguel Palhas 244e8d2927 Merge structured-view/185-lock-order into blitz/structured-view 2026-09-05 05:20:20 +01:00
Miguel Palhas 4a1a10c49a fix(maestro): avoid nested session locks 2026-09-05 05:15:58 +01:00
Miguel Palhas d607740fea Merge structured-view/184-approval-persist into blitz/structured-view 2026-09-05 05:05:15 +01:00
Miguel PalhasandClaude Sonnet 5 0b34d1608f fix(maestro): make FaultyLogSink honor the Write contract on partial writes
write() reported Ok(partial) and Err from the same call, which Write
forbids. Defer the error to the following call so write_all observes
a genuine short write, then a failure — matching what a real fd does.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
2026-09-05 05:02:01 +01:00
Miguel Palhas 075dea509a fix(maestro): make EventLog::append transactional under partial write/flush failure
A write_all that errors mid-line, or a flush that errors after a full
write, could leave last_seq and the log's on-disk line count out of
sync: a well-formed but uncounted line, or a partial one, would corrupt
every later cursor once the next append landed beside it. append now
captures the pre-append length and truncates back to it on either
failure, so the same EventLog can be appended to again immediately. If
the rollback itself can't restore that length, append returns a Fatal
variant so callers kill the session instead of promising consistency
they can no longer keep — record_permission_request and its pump now
route this through the session-death path used elsewhere.

Adds deterministic fault injection (a seekable/truncatable LogSink
double) to prove the same EventLog recovers with valid JSONL and
live/replay agreement, rather than swapping in a healthy handle. Fixes
the row-index rollback test to actually shift a later row, and
tightens the persistence-failure wording in STRUCTURED-VIEW.md.
2026-09-05 04:55:26 +01:00
Miguel Palhas c067f48a24 fix(maestro): roll back phantom approval row on log append failure
request_approval() appended the Approval row to the live projection
before the daemon tried to persist it. On append failure the pending
nonce was cancelled but the row stayed in the in-memory transcript, so
a WebSocket snapshot taken afterward showed a row durable replay could
never reproduce.

Add Projection::rollback_row to undo the append and call it alongside
the existing cancel_pending_approvals on failure. Extract the pump's
per-request handling into record_permission_request so the failure
path is directly testable without a live ACP connection.

Fault-injection regression tests use a read-only file handle
(EventLog::for_test, cfg(test)) to force append() to fail
deterministically, then assert the live snapshot and durable replay
agree (both empty) and a subsequent request still persists and
replays correctly.

Fixes #184.
2026-09-05 04:34:47 +01:00
Miguel Palhas 51f96cfb98 Merge structured-view/142-structured-activity into blitz/structured-view 2026-09-05 04:15:46 +01:00
Miguel Palhas 8f5d96a95b feat(maestro): derive structured activity 2026-09-05 04:08:43 +01:00
Miguel Palhas 3e98b3ab24 Merge structured-view/137-structured-docs into blitz/structured-view 2026-09-05 03:51:34 +01:00
Miguel Palhas 24ed1ac002 docs(maestro): fix two more factual errors in structured docs
INSTALL.md conflated event-log replay with ACP session/load resume.
STRUCTURED-VIEW.md overstated death-on-append-failure: only notification
appends kill the session, approval-row append failures just cancel the
pending approval.
2026-09-05 03:49:59 +01:00
Miguel PalhasandClaude Sonnet 5 64741a027e docs(maestro): fix five more factual errors in structured-session docs
Corrects recovery-as-two-operations, tmux-only survival on restart,
replay's durable-vs-live-state gap, snapshot-vs-replay purpose, and the
three-file persistence count.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
2026-09-05 03:45:26 +01:00
Miguel Palhas cdd4213ff5 docs(maestro): fix three factual errors in STRUCTURED-VIEW.md
- replay: since=0 returns the first page (still limit-bounded), not the
  whole log; clients page forward from the last returned cursor
- PR_SET_PDEATHSIG is Linux-only; other platforms depend on the
  terminate() ladder running to completion
- approval nonce is minted before cancelling the prior pending approval,
  not after
2026-09-05 03:38:32 +01:00
Miguel Palhas f8973b3802 docs(maestro): write STRUCTURED-VIEW.md
Contracts the eight transcript row kinds, daemon-vs-client decisions, the
transcript/reduced-state projections and their cursor/replay semantics, the
ACP handshake and declined capabilities, the adapter install command, and
the approval nonce rules (generation, live-only placement, secrecy from the
agent and durable history, single-use/duplicate resolution) and bounded
prompt/cancel liveness.

Amends RESTART.md to state plainly that a structured session's agent is a
daemon child, not tmux-backed: a restart ends the active turn, the
transcript survives in the event log, and pending approvals/nonces do not.

Refs #137
2026-09-05 03:32:42 +01:00
Miguel Palhas 1e2c0f0a38 Merge structured-view/139-permission-requests into blitz/structured-view 2026-09-05 03:14:19 +01:00
Miguel Palhas 6fa7dc63e4 fix: serialize structured prompt dispatch 2026-09-05 03:11:16 +01:00
Miguel Palhas 5c0a4c4640 fix: bind approvals to prompt turns 2026-09-05 02:57:29 +01:00
Miguel Palhas 8773671afd fix(daemon): unify approval replay log 2026-09-05 02:44:59 +01:00
Miguel Palhas d9815d86c6 feat(daemon): resolve ACP approvals 2026-09-05 02:32:01 +01:00
Miguel Palhas 13009d0f0e Merge structured-view/138-structured-api into blitz/structured-view 2026-09-05 02:16:00 +01:00
Miguel Palhas bbae09bc48 fix(daemon): reject dead structured sessions 2026-09-05 02:13:55 +01:00
Miguel Palhas c7c294996e fix(daemon): remove unused turn token 2026-09-05 02:07:07 +01:00
Miguel Palhas 3c638b08de fix: clarify structured prompt failures 2026-09-05 02:05:06 +01:00
Miguel Palhas be23a2b539 fix: harden structured replay 2026-09-05 01:56:07 +01:00
Miguel Palhas b98f5cb3e5 feat(maestro): serve structured session API 2026-09-05 01:46:21 +01:00
Miguel Palhas 0f5576c77a Merge structured-view/136-transcript-projection into blitz/structured-view 2026-09-05 01:31:55 +01:00
17 changed files with 3740 additions and 136 deletions
+6
View File
@@ -92,6 +92,12 @@ clients — how it opens, the sigils, the key model, the match-and-rank algorith
the two matchers must agree on, and the fixture table they are tested against.
Read it before changing either bar or either matcher.
`docs/STRUCTURED-VIEW.md` is the contract for a structured session (an agent
run as a daemon child over ACP instead of in a tmux pane) — the eight
transcript row kinds, the two projections and their cursor/replay semantics,
the handshake and declined capabilities, and the approval nonce rules. Read
it before changing anything that renders a structured session.
`nix run .#dev` (or `just dev`) brings up the whole local stack under
process-compose: a debug daemon on 127.0.0.1:7071, then Vite on :5173 once the
daemon answers `/api/info`. The SPA proxies `/api/*` back to the daemon, so only
+139 -12
View File
@@ -42,9 +42,10 @@ use std::sync::{Arc, Mutex};
use std::time::Duration;
use agent_client_protocol::schema::v1::{
CancelNotification, ClientCapabilities, FileSystemCapabilities, Implementation,
InitializeRequest, LoadSessionRequest, NewSessionRequest, RequestPermissionOutcome,
RequestPermissionRequest, RequestPermissionResponse, SessionId, SessionNotification,
CancelNotification, ClientCapabilities, ContentBlock, FileSystemCapabilities, Implementation,
InitializeRequest, LoadSessionRequest, NewSessionRequest, PromptRequest,
RequestPermissionOutcome, RequestPermissionRequest, RequestPermissionResponse, SessionId,
SessionNotification,
};
use agent_client_protocol::schema::ProtocolVersion;
use agent_client_protocol::{ByteStreams, Client};
@@ -112,6 +113,15 @@ impl SpawnConfig {
}
}
/// A blocked ACP permission request delivered to the daemon.
#[derive(Debug)]
pub struct PermissionRequest {
/// The prompt turn that was active when ACP delivered this callback.
pub turn: u64,
pub request: RequestPermissionRequest,
pub response: oneshot::Sender<RequestPermissionOutcome>,
}
#[derive(Debug, thiserror::Error)]
pub enum AcpError {
/// The adapter is not on `PATH`. The message names the install command
@@ -153,6 +163,11 @@ fn is_executable(path: &Path) -> bool {
enum Command {
Cancel,
Prompt {
text: String,
turn: u64,
done: oneshot::Sender<Result<(), String>>,
},
Close,
}
@@ -161,6 +176,31 @@ enum ChildCommand {
Reap(oneshot::Sender<bool>),
}
#[derive(Debug, Default)]
struct PermissionTurnState {
active: Option<u64>,
}
impl PermissionTurnState {
fn begin(&mut self, turn: u64) {
self.active = Some(turn);
}
fn cancel_active(&mut self) {
self.active = None;
}
fn finish(&mut self, turn: u64) {
if self.active == Some(turn) {
self.active = None;
}
}
fn callback_turn(&self) -> Option<u64> {
self.active
}
}
/// Why the handshake did not produce a session id.
enum ReadyError {
/// The adapter answered with an error.
@@ -187,6 +227,7 @@ pub struct AcpSession {
commands: mpsc::UnboundedSender<Command>,
child_commands: mpsc::UnboundedSender<ChildCommand>,
updates: Mutex<Option<mpsc::UnboundedReceiver<SessionNotification>>>,
permission_requests: Mutex<Option<mpsc::UnboundedReceiver<PermissionRequest>>>,
exit: watch::Receiver<Option<ExitStatus>>,
}
@@ -230,12 +271,36 @@ impl AcpSession {
self.updates.lock().expect("updates mutex").take()
}
pub fn take_permission_requests(&self) -> Option<mpsc::UnboundedReceiver<PermissionRequest>> {
self.permission_requests
.lock()
.expect("permission requests")
.take()
}
/// Send `session/cancel`. Fire-and-forget: a closed connection means the
/// child is already gone, which is the outcome cancel was after.
pub fn cancel(&self) {
let _ = self.commands.send(Command::Cancel);
}
/// Send a user prompt and receive completion once the agent ends the turn.
pub fn prompt(
&self,
text: String,
turn: u64,
) -> Result<oneshot::Receiver<Result<(), String>>, String> {
let (done_tx, done_rx) = oneshot::channel();
self.commands
.send(Command::Prompt {
text,
turn,
done: done_tx,
})
.map_err(|_| "ACP connection is closed".to_string())?;
Ok(done_rx)
}
/// End the child: `session/cancel`, wait `cancel_grace`, `SIGTERM`, wait
/// `term_grace`, `SIGKILL`. Returns once the child has been reaped.
pub async fn terminate(&self, cancel_grace: Duration, term_grace: Duration) {
@@ -318,6 +383,7 @@ pub async fn spawn(config: SpawnConfig) -> Result<AcpSession, AcpError> {
tokio::spawn(supervise_child(child, pid, child_commands_rx, exit_tx));
let (updates_tx, updates_rx) = mpsc::unbounded_channel();
let (permission_tx, permission_rx) = mpsc::unbounded_channel();
let (commands_tx, commands_rx) = mpsc::unbounded_channel();
let (ready_tx, ready_rx) = oneshot::channel();
let transport = ByteStreams::new(stdin.compat_write(), stdout.compat());
@@ -326,6 +392,7 @@ pub async fn spawn(config: SpawnConfig) -> Result<AcpSession, AcpError> {
config.cwd.clone(),
config.load_session.clone(),
updates_tx,
permission_tx,
commands_rx,
ready_tx,
));
@@ -339,6 +406,7 @@ pub async fn spawn(config: SpawnConfig) -> Result<AcpSession, AcpError> {
commands: commands_tx,
child_commands: child_commands_tx,
updates: Mutex::new(Some(updates_rx)),
permission_requests: Mutex::new(Some(permission_rx)),
exit: exit_rx,
});
}
@@ -450,6 +518,7 @@ async fn run_connection<T>(
cwd: PathBuf,
load_session: Option<String>,
updates_tx: mpsc::UnboundedSender<SessionNotification>,
permission_requests: mpsc::UnboundedSender<PermissionRequest>,
mut commands_rx: mpsc::UnboundedReceiver<Command>,
ready_tx: oneshot::Sender<Result<String, ReadyError>>,
) where
@@ -461,6 +530,10 @@ async fn run_connection<T>(
// history the event log already holds (#135).
let replaying = Arc::new(AtomicBool::new(false));
let replaying_seen = replaying.clone();
// ACP callbacks have no turn id. The daemon never queues a later prompt
// until a cancelled prompt settles, so this active turn is unambiguous.
let permission_turn = Arc::new(Mutex::new(PermissionTurnState::default()));
let permission_turn_seen = permission_turn.clone();
let result = Client
.builder()
.name("maestro")
@@ -478,15 +551,30 @@ async fn run_connection<T>(
)
.on_receive_request(
async move |request: RequestPermissionRequest, responder, _cx| {
// Stub until #139: decline rather than grant something the
// operator never saw.
tracing::warn!(
session = %request.session_id,
"session/request_permission declined: not implemented yet (#139)"
);
responder.respond(RequestPermissionResponse::new(
RequestPermissionOutcome::Cancelled,
))
let turn = permission_turn_seen
.lock()
.expect("permission turn state")
.callback_turn();
let Some(turn) = turn else {
return responder.respond(RequestPermissionResponse::new(
RequestPermissionOutcome::Cancelled,
));
};
let (response, received) = oneshot::channel();
if permission_requests
.send(PermissionRequest {
turn,
request,
response,
})
.is_err()
{
return responder.respond(RequestPermissionResponse::new(
RequestPermissionOutcome::Cancelled,
));
}
let outcome = received.await.unwrap_or(RequestPermissionOutcome::Cancelled);
responder.respond(RequestPermissionResponse::new(outcome))
},
agent_client_protocol::on_receive_request!(),
)
@@ -558,10 +646,49 @@ async fn run_connection<T>(
tokio::select! {
cmd = commands_rx.recv() => match cmd {
Some(Command::Cancel) => {
permission_turn
.lock()
.expect("permission turn state")
.cancel_active();
let _ = cx.send_notification(CancelNotification::new(
SessionId::new(session_id.as_str()),
));
}
Some(Command::Prompt { text, turn, done }) => {
permission_turn
.lock()
.expect("permission turn state")
.begin(turn);
let request = PromptRequest::new(
SessionId::new(session_id.as_str()),
vec![ContentBlock::from(text)],
);
let done = Arc::new(Mutex::new(Some(done)));
let callback_done = done.clone();
let callback_turn = permission_turn.clone();
if let Err(error) = cx.send_request(request).on_receiving_result(move |result| {
let callback_done = callback_done.clone();
let callback_turn = callback_turn.clone();
async move {
callback_turn
.lock()
.expect("permission turn state")
.finish(turn);
if let Some(done) = callback_done.lock().expect("prompt sender").take() {
let _ = done.send(result.map(|_| ()).map_err(|e| e.to_string()));
}
Ok(())
}
}) {
permission_turn
.lock()
.expect("permission turn state")
.finish(turn);
if let Some(done) = done.lock().expect("prompt sender").take() {
let _ = done.send(Err(error.to_string()));
}
}
}
Some(Command::Close) | None => break,
},
_ = cx.incoming_closed() => break,
@@ -24,11 +24,18 @@ Environment knobs, all optional:
so the SIGTERM path is exercised)
FAKE_ACP_IGNORE_EOF=1 keep running after stdin closes, until signalled
FAKE_ACP_EMIT_ON_CANCEL=1 emit a final notification before handling cancel
FAKE_ACP_CANCEL_SETTLES=1 complete the outstanding prompt on cancel
FAKE_ACP_NO_LOAD=1 advertise `loadSession: false`
FAKE_ACP_LOAD_REPLAY=n on `session/load`, replay n `agent_message_chunk`
notifications ("replayed <i>") before answering,
the way a real agent streams history back
FAKE_ACP_LOAD_FAIL=1 answer `session/load` with a JSON-RPC error
FAKE_ACP_DELAY_PROMPT_MS=n wait before answering `session/prompt`
FAKE_ACP_PROMPT_FAIL=1 answer every `session/prompt` with a JSON-RPC error
FAKE_ACP_PROMPT_FAIL_ONCE=1 answer only the first `session/prompt` with a
JSON-RPC error
FAKE_ACP_PERMISSION_AFTER_CANCEL=1 defer the first prompt's permission
callback until a later prompt
Signals: SIGUSR1 emits one `agent_message_chunk` notification ("tick <n>",
counting from 1) and logs `emit tick <n>`, so a test can make the agent
@@ -65,6 +72,10 @@ signal.signal(signal.SIGTERM, on_sigterm)
session_id = f"fake-{os.getpid()}"
ticks = 0
prompt_failures = 0
pending_prompt_id = None
pending_permission_id = None
delayed_permission = False
def chunk(text):
@@ -97,6 +108,12 @@ for raw in sys.stdin:
method = msg.get("method")
if method is None:
# A response to something we sent; nothing to do.
if msg.get("id") == pending_permission_id:
log("permission response " + json.dumps(msg.get("result"), sort_keys=True))
send({"jsonrpc": "2.0", "id": pending_prompt_id,
"result": {"stopReason": "end_turn"}})
pending_prompt_id = None
pending_permission_id = None
continue
log(f"recv {method}")
rid = msg.get("id")
@@ -135,8 +152,56 @@ for raw in sys.stdin:
elif method == "session/cancel":
if os.environ.get("FAKE_ACP_EMIT_ON_CANCEL") == "1":
send(chunk("final on cancel"))
if (os.environ.get("FAKE_ACP_CANCEL_SETTLES") == "1"
and pending_prompt_id is not None):
send({"jsonrpc": "2.0", "id": pending_prompt_id,
"result": {"stopReason": "cancelled"}})
pending_prompt_id = None
pending_permission_id = None
if os.environ.get("FAKE_ACP_EXIT_ON_CANCEL") == "1":
sys.exit(0)
elif method == "session/prompt":
if os.environ.get("FAKE_ACP_PERMISSION_AFTER_CANCEL") == "1":
if pending_prompt_id is None:
pending_prompt_id = rid
delayed_permission = True
continue
if delayed_permission:
delayed_permission = False
pending_permission_id = "permission-1"
send({"jsonrpc": "2.0", "id": pending_permission_id,
"method": "session/request_permission", "params": {
"sessionId": session_id,
"toolCall": {"toolCallId": "tool-1", "fields": {}},
"options": [
{"optionId": "once", "name": "Allow once", "kind": "allow_once"},
{"optionId": "deny", "name": "Deny", "kind": "reject_once"}
]
}})
continue
if os.environ.get("FAKE_ACP_REQUEST_PERMISSION") == "1":
pending_prompt_id = rid
pending_permission_id = "permission-1"
send({"jsonrpc": "2.0", "id": pending_permission_id,
"method": "session/request_permission", "params": {
"sessionId": session_id,
"toolCall": {"toolCallId": "tool-1", "fields": {}},
"options": [
{"optionId": "once", "name": "Allow once", "kind": "allow_once"},
{"optionId": "session", "name": "Allow session", "kind": "allow_always"},
{"optionId": "deny", "name": "Deny", "kind": "reject_once"}
]
}})
continue
time.sleep(int(os.environ.get("FAKE_ACP_DELAY_PROMPT_MS", "0")) / 1000)
fail_once = os.environ.get("FAKE_ACP_PROMPT_FAIL_ONCE") == "1"
if os.environ.get("FAKE_ACP_PROMPT_FAIL") == "1" or (fail_once and prompt_failures == 0):
prompt_failures += 1
send({"jsonrpc": "2.0", "id": rid,
"error": {"code": -32000, "message": "fake prompt failure"}})
else:
send({"jsonrpc": "2.0", "id": rid,
"result": {"stopReason": "end_turn"}})
elif rid is not None:
send({"jsonrpc": "2.0", "id": rid,
"error": {"code": -32601, "message": f"method not found: {method}"}})
+28
View File
@@ -66,6 +66,23 @@ pub fn decide(idle_for: Duration, awaiting_input: bool) -> ActivityState {
}
}
/// Derive a structured session's sidebar state from its ACP projection.
pub fn decide_structured(
turn_active: bool,
pending_approval: bool,
turn_failed: bool,
) -> ActivityState {
if turn_failed {
ActivityState::Error
} else if pending_approval {
ActivityState::AwaitingInput
} else if turn_active {
ActivityState::Active
} else {
ActivityState::Idle
}
}
/// How long a session's pane must be continuously focused (a client's pane,
/// not a sidebar highlight) before an `AwaitingInput` glow clears: the
/// operator has looked, so the rail stops asking for attention.
@@ -136,6 +153,17 @@ mod tests {
);
}
#[test]
fn structured_activity_follows_lifecycle_with_approval_precedence() {
assert_eq!(decide_structured(false, false, false), ActivityState::Idle);
assert_eq!(decide_structured(true, false, false), ActivityState::Active);
assert_eq!(
decide_structured(true, true, false),
ActivityState::AwaitingInput
);
assert_eq!(decide_structured(false, false, true), ActivityState::Error);
}
#[test]
fn report_is_in_force_until_exactly_its_expiry() {
let ttl = Duration::from_secs(300);
+205 -8
View File
@@ -21,12 +21,13 @@ use maestro_config::SpawnPreset;
use maestro_protocol::{
routes, ApiError, CatalogEvent, CreateSessionRequest, DaemonInfo, GitRoot, NameBatch,
RenameSessionRequest, ReportActivityRequest, SendInputRequest, Session, SetFocusRequest,
PROTOCOL_VERSION,
StructuredApprovalDecision, StructuredEvent, StructuredPromptRequest, PROTOCOL_VERSION,
};
use serde::Deserialize;
use uuid::Uuid;
use crate::session::SessionManager;
use crate::session::{QueueAcpPrompt, SessionManager};
use crate::sidecar::EventCursor;
#[derive(Debug, Clone)]
pub struct LocalApiState {
@@ -64,7 +65,15 @@ pub fn router(state: LocalApiState) -> Router {
.route(routes::NAMES, get(names))
.route(routes::INFO, get(daemon_info))
.route(routes::EVENTS, any(events_ws))
.route(routes::patterns::ATTACH, any(attach_ws));
.route(routes::patterns::ATTACH, any(attach_ws))
.route(routes::patterns::STRUCTURED, any(structured_ws))
.route(routes::patterns::STRUCTURED_REPLAY, get(structured_replay))
.route(routes::patterns::STRUCTURED_PROMPT, post(structured_prompt))
.route(routes::patterns::STRUCTURED_CANCEL, post(structured_cancel))
.route(
routes::patterns::STRUCTURED_APPROVAL,
post(structured_approval),
);
#[cfg(feature = "haiku-autoname")]
{
router = router.route(routes::patterns::SESSION_PROMPT, post(session_prompt));
@@ -298,11 +307,10 @@ async fn rename_session(
}
}
/// `POST /api/sessions/:id/activity` (#105): an agent reports its own state.
/// The report holds above the daemon's inference for a TTL, then inference
/// resumes — an agent that dies mid-run cannot pin its session at `Error`
/// forever. Unknown session is 404; an unknown state value is rejected by the
/// JSON decoder (422) rather than coerced. Answers `204 No Content`.
/// `POST /api/sessions/:id/activity` (#105): an agent reports a terminal
/// session's state. Structured sessions acknowledge and ignore these reports
/// because their activity comes from ACP. Unknown session is 404; an unknown
/// state value is rejected by the JSON decoder (422). Answers `204 No Content`.
async fn report_activity(
State(s): State<LocalApiState>,
Path(id): Path<Uuid>,
@@ -369,6 +377,126 @@ fn error_response(status: StatusCode, code: &str, message: &str) -> Response {
(status, Json(ApiError::new(code, message))).into_response()
}
#[derive(Debug, Deserialize)]
struct ReplayQuery {
since: Option<u64>,
before: Option<u64>,
limit: Option<usize>,
}
const MAX_STRUCTURED_REPLAY_LIMIT: usize = 500;
async fn structured_replay(
State(s): State<LocalApiState>,
Path(id): Path<Uuid>,
Query(query): Query<ReplayQuery>,
) -> Response {
if query.since.is_some() && query.before.is_some() {
return error_response(
StatusCode::BAD_REQUEST,
"invalid_cursor",
"use either since or before, not both",
);
}
let cursor = match query.before {
Some(before) => EventCursor::Before(Some(before)),
None => EventCursor::Since(query.since.unwrap_or(0)),
};
match s
.manager
.read_acp_projection_replay(
id,
cursor,
query
.limit
.unwrap_or(100)
.clamp(1, MAX_STRUCTURED_REPLAY_LIMIT),
)
.await
{
Ok(Some(events)) => Json(events).into_response(),
// A terminal session, or a structured session whose adapter is no
// longer live, cannot provide a coherent replay continuation.
Ok(None) => error_response(
StatusCode::CONFLICT,
"not_structured",
"session is not a live structured session",
),
Err(e) => {
let info = e.to_error_info();
error_response(session_error_status(&info.code), &info.code, &info.message)
}
}
}
async fn structured_prompt(
State(s): State<LocalApiState>,
Path(id): Path<Uuid>,
Json(body): Json<StructuredPromptRequest>,
) -> Response {
match s.manager.queue_acp_prompt(id, body.prompt).await {
Ok(QueueAcpPrompt::Queued) => StatusCode::NO_CONTENT.into_response(),
Ok(QueueAcpPrompt::InFlight) => error_response(
StatusCode::CONFLICT,
"turn_in_flight",
"the structured session already has an active turn",
),
Ok(QueueAcpPrompt::Unavailable) => error_response(
StatusCode::CONFLICT,
"not_structured",
"session is not a live structured session",
),
Ok(QueueAcpPrompt::Failed(message)) => {
error_response(StatusCode::BAD_GATEWAY, "acp_prompt_failed", &message)
}
Err(e) => {
let info = e.to_error_info();
error_response(session_error_status(&info.code), &info.code, &info.message)
}
}
}
async fn structured_cancel(State(s): State<LocalApiState>, Path(id): Path<Uuid>) -> Response {
match s.manager.cancel_acp_turn(id).await {
Ok(true) => StatusCode::NO_CONTENT.into_response(),
Ok(false) => error_response(
StatusCode::CONFLICT,
"not_structured",
"session is not a live structured session",
),
Err(e) => {
let info = e.to_error_info();
error_response(session_error_status(&info.code), &info.code, &info.message)
}
}
}
async fn structured_approval(
State(s): State<LocalApiState>,
Path((id, nonce)): Path<(Uuid, String)>,
Json(decision): Json<StructuredApprovalDecision>,
) -> Response {
let decision = match decision {
StructuredApprovalDecision::AllowOnce => crate::session::ApprovalDecision::AllowOnce,
StructuredApprovalDecision::AllowForSession => {
crate::session::ApprovalDecision::AllowForSession
}
StructuredApprovalDecision::Deny => crate::session::ApprovalDecision::Deny,
};
match s.manager.resolve_acp_approval(id, &nonce, decision).await {
Ok(Some(())) => StatusCode::NO_CONTENT.into_response(),
Ok(None) => error_response(
StatusCode::CONFLICT,
"not_structured",
"session is not a live structured session",
),
Err(e) => {
let info = e.to_error_info();
error_response(session_error_status(&info.code), &info.code, &info.message)
}
}
}
/// Convenience constructor — pulls the bits `LocalApiState` needs from
/// `DaemonConfig` so callers don't have to assemble it by hand. The project
/// scanner starts with its cache cold; wire
@@ -419,6 +547,75 @@ async fn attach_ws(
}))
}
#[allow(clippy::result_large_err)]
async fn structured_ws(
Path(id): Path<Uuid>,
State(s): State<LocalApiState>,
ws: axum::extract::WebSocketUpgrade,
) -> Result<Response, Response> {
let Some((snapshot, state, rx)) = s
.manager
.subscribe_acp_projection(id, EventCursor::Since(0))
.await
.map_err(|e| {
let info = e.to_error_info();
error_response(session_error_status(&info.code), &info.code, &info.message)
})?
else {
return Err(error_response(
StatusCode::CONFLICT,
"not_structured",
"session is not a live structured session",
));
};
Ok(ws.on_upgrade(move |socket| async move {
let _ = pump_structured(socket, snapshot, state, rx).await;
}))
}
async fn pump_structured(
socket: axum::extract::ws::WebSocket,
snapshot: maestro_protocol::TranscriptSnapshot,
state: maestro_protocol::ReducedState,
mut rx: tokio::sync::broadcast::Receiver<crate::projection::ProjectionUpdate>,
) -> Result<(), axum::Error> {
use axum::extract::ws::Message;
use futures_util::{SinkExt, StreamExt};
let (mut sink, mut stream) = socket.split();
for event in [
StructuredEvent::TranscriptSnapshot { snapshot },
StructuredEvent::ReducedState { state },
] {
if sink
.send(Message::Text(
serde_json::to_string(&event).expect("structured event encode"),
))
.await
.is_err()
{
return Ok(());
}
}
loop {
tokio::select! {
update = rx.recv() => match update {
Ok(update) => {
let event = StructuredEvent::ProjectionUpdate { transcript: update.transcript, state: update.state };
if sink.send(Message::Text(serde_json::to_string(&event).expect("structured event encode"))).await.is_err() { break; }
}
Err(tokio::sync::broadcast::error::RecvError::Lagged(_)) => break,
Err(tokio::sync::broadcast::error::RecvError::Closed) => break,
},
frame = stream.next() => match frame {
None | Some(Err(_)) | Some(Ok(Message::Close(_))) => break,
Some(Ok(_)) => {},
},
}
}
Ok(())
}
// ---------------------------------------------------------------------------
// WS /api/events — live catalog deltas for browser/TUI subscribers (NAP-248)
// ---------------------------------------------------------------------------
+282 -30
View File
@@ -4,7 +4,8 @@ use std::collections::{HashMap, HashSet};
use maestro_acp::schema::{ContentChunk, SessionNotification, SessionUpdate, ToolCall};
use maestro_protocol::{
ReducedState, TranscriptDelta, TranscriptRow, TranscriptRowKind, TranscriptSnapshot,
ActivityState, ReducedState, TranscriptDelta, TranscriptRow, TranscriptRowKind,
TranscriptSnapshot,
};
use serde_json::Value;
@@ -35,7 +36,9 @@ pub struct Projection {
tools: HashMap<String, ToolCall>,
pending_tool_updates: HashMap<String, Vec<maestro_acp::schema::ToolCallUpdateFields>>,
state: ReducedState,
turn_active: bool,
active_turn: Option<u64>,
turn_failed: bool,
next_turn: u64,
turn_tool_calls: HashSet<String>,
}
@@ -104,27 +107,134 @@ impl Projection {
self.state.clone()
}
/// Apply a transport-owned prompt lifecycle transition.
pub fn turn_lifecycle(&mut self, lifecycle: TurnLifecycle) -> ProjectionUpdate {
match lifecycle {
TurnLifecycle::Begin => {
self.turn_active = true;
self.state.turn.responding = true;
self.state.turn.thinking = true;
self.turn_tool_calls.clear();
self.state.turn.active_tool_calls.clear();
}
TurnLifecycle::Finish | TurnLifecycle::Cancel | TurnLifecycle::Failure => {
self.turn_active = false;
self.state.turn.responding = false;
self.state.turn.thinking = false;
self.turn_tool_calls.clear();
self.state.turn.active_tool_calls.clear();
}
pub fn turn_active(&self) -> bool {
self.active_turn.is_some()
}
/// Sidebar activity derived from the live turn projection.
pub fn activity(&self) -> ActivityState {
crate::activity::decide_structured(
self.active_turn.is_some(),
self.state.pending_approval.is_some(),
self.turn_failed,
)
}
/// Start a prompt turn and return its monotonically increasing token.
pub fn begin_turn(&mut self) -> (u64, ProjectionUpdate) {
self.next_turn += 1;
let turn = self.next_turn;
self.active_turn = Some(turn);
self.turn_failed = false;
self.state.turn.responding = true;
self.state.turn.thinking = true;
self.turn_tool_calls.clear();
self.state.turn.active_tool_calls.clear();
(
turn,
ProjectionUpdate {
transcript: None,
state: self.state(),
},
)
}
/// End `turn` only when it remains the active prompt turn.
pub fn turn_lifecycle(
&mut self,
lifecycle: TurnLifecycle,
turn: u64,
) -> Option<ProjectionUpdate> {
if !matches!(
lifecycle,
TurnLifecycle::Finish | TurnLifecycle::Cancel | TurnLifecycle::Failure
) || self.active_turn != Some(turn)
{
return None;
}
ProjectionUpdate {
self.active_turn = None;
self.turn_failed = lifecycle == TurnLifecycle::Failure;
self.state.turn.responding = false;
self.state.turn.thinking = false;
self.turn_tool_calls.clear();
self.state.turn.active_tool_calls.clear();
Some(ProjectionUpdate {
transcript: None,
state: self.state(),
})
}
pub fn active_turn(&self) -> Option<u64> {
self.active_turn
}
/// Add a daemon-originated permission row. The nonce belongs only in the
/// reduced state, which is the client control surface, never in ACP data.
pub fn request_approval(&mut self, nonce: String, request: Value) -> ProjectionUpdate {
let row = self.append(
format!("approval:{}", uuid::Uuid::new_v4()),
TranscriptRowKind::Approval,
request.clone(),
);
self.state.pending_approval = Some(serde_json::json!({
"nonce": nonce,
"request": request,
}));
ProjectionUpdate {
transcript: Some(row),
state: self.state(),
}
}
pub fn restore_approval(&mut self, row: TranscriptRow) -> ProjectionUpdate {
let transcript = if !self.row_index.contains_key(&row.id) {
self.row_index.insert(row.id.clone(), self.rows.len());
self.rows.push(row.clone());
Some(TranscriptDelta::Append { row })
} else {
None
};
ProjectionUpdate {
transcript,
state: self.state(),
}
}
/// Clear an approval only when it is still the one the client answered.
pub fn resolve_approval(&mut self, nonce: &str) -> Option<ProjectionUpdate> {
let pending = self.state.pending_approval.as_ref()?;
if pending.get("nonce").and_then(Value::as_str) != Some(nonce) {
return None;
}
self.state.pending_approval = None;
Some(ProjectionUpdate {
transcript: None,
state: self.state(),
})
}
/// Terminal turn paths must not leave a recovered snapshot permanently
/// blocked on an ACP request whose transport no longer exists.
pub fn clear_approval(&mut self) -> Option<ProjectionUpdate> {
self.state.pending_approval.take()?;
Some(ProjectionUpdate {
transcript: None,
state: self.state(),
})
}
/// Undo an `append`ed row that a caller decides never happened — used
/// when persisting the row's own event fails (#184), so a snapshot taken
/// after the failure cannot show a row durable replay would never
/// reproduce. Only ever called for a row that was just appended and
/// never patched, so removing it by id fully reverses `append`.
pub fn rollback_row(&mut self, id: &str) {
let Some(index) = self.row_index.remove(id) else {
return;
};
self.rows.remove(index);
for (shifted_index, row) in self.rows.iter().enumerate().skip(index) {
self.row_index.insert(row.id.clone(), shifted_index);
}
}
@@ -230,28 +340,28 @@ impl Projection {
}
fn agent_text(&mut self) {
if self.turn_active {
if self.turn_active() {
self.state.turn.responding = true;
self.state.turn.thinking = false;
}
}
fn agent_thought(&mut self) {
if self.turn_active {
if self.turn_active() {
self.state.turn.responding = true;
self.state.turn.thinking = true;
}
}
fn track_tool_call(&mut self, tool: &ToolCall) {
if self.turn_active {
if self.turn_active() {
self.turn_tool_calls.insert(tool.tool_call_id.to_string());
self.sync_active_tool(&tool.tool_call_id.to_string());
}
}
fn sync_active_tool(&mut self, tool_id: &str) {
if !self.turn_active || !self.turn_tool_calls.contains(tool_id) {
if !self.turn_active() || !self.turn_tool_calls.contains(tool_id) {
return;
}
let active = self.tools.get(tool_id).is_some_and(|tool| {
@@ -328,6 +438,16 @@ mod tests {
.unwrap()
}
#[test]
fn a_session_update_tag_the_acp_dependency_does_not_know_fails_to_deserialize() {
// `SessionUpdate` has no unknown-variant fallback, so this never reaches `Projection::apply`.
let result: Result<SessionNotification, _> = serde_json::from_value(serde_json::json!({
"sessionId": "acp-1",
"update": {"sessionUpdate": "some_future_update_variant"},
}));
assert!(result.is_err());
}
#[test]
fn folds_content_tool_diff_plan_and_notice_rows() {
let mut projection = Projection::default();
@@ -538,7 +658,7 @@ mod tests {
);
assert_eq!(projection.state().turn, Default::default());
projection.turn_lifecycle(TurnLifecycle::Begin);
let (_, _) = projection.begin_turn();
assert!(projection.state().turn.responding);
assert!(projection.state().turn.thinking);
projection.apply(
@@ -558,8 +678,8 @@ mod tests {
TurnLifecycle::Cancel,
TurnLifecycle::Failure,
] {
projection.turn_lifecycle(TurnLifecycle::Begin);
projection.turn_lifecycle(terminal);
let (turn, _) = projection.begin_turn();
projection.turn_lifecycle(terminal, turn);
assert!(!projection.state().turn.responding);
assert!(!projection.state().turn.thinking);
}
@@ -571,7 +691,7 @@ mod tests {
next_id: &str,
) {
let mut projection = Projection::default();
projection.turn_lifecycle(TurnLifecycle::Begin);
let (turn, _) = projection.begin_turn();
projection.apply(
1,
&event(serde_json::json!({
@@ -581,11 +701,11 @@ mod tests {
);
assert_eq!(projection.state().turn.active_tool_calls, [unfinished_id]);
projection.turn_lifecycle(terminal);
projection.turn_lifecycle(terminal, turn);
assert!(projection.state().turn.active_tool_calls.is_empty());
assert_eq!(projection.snapshot(1).rows.len(), 1);
projection.turn_lifecycle(TurnLifecycle::Begin);
projection.begin_turn();
projection.apply(
2,
&event(serde_json::json!({
@@ -605,4 +725,136 @@ mod tests {
fn failed_turn_does_not_reactivate_unfinished_tools() {
terminal_turn_drops_unfinished_tools(TurnLifecycle::Failure, "failed-1", "next-1");
}
#[test]
fn approval_history_survives_pending_cleanup() {
let mut projection = Projection::default();
let update =
projection.request_approval("secret".into(), serde_json::json!({"tool":"bash"}));
let TranscriptDelta::Append { row } = update.transcript.unwrap() else {
panic!("approval row")
};
assert_eq!(row.kind, TranscriptRowKind::Approval);
assert!(row.data.get("nonce").is_none());
assert!(!row.id.contains("secret"));
let persisted =
serde_json::to_string(&crate::sidecar::PersistedEvent::Approval { row: row.clone() })
.unwrap();
assert!(!persisted.contains("secret"));
assert_eq!(
projection.state().pending_approval.unwrap()["nonce"],
"secret"
);
projection.clear_approval();
assert!(projection.state().pending_approval.is_none());
assert_eq!(projection.snapshot(0).rows, vec![row]);
}
#[test]
fn rollback_row_removes_only_that_row_and_keeps_others_addressable() {
let mut projection = Projection::default();
projection.apply(
1,
&event(serde_json::json!({
"sessionUpdate": "user_message_chunk",
"messageId": "m1",
"content": {"type": "text", "text": "hi"},
})),
);
let TranscriptDelta::Append { row } = projection
.request_approval("nonce".into(), serde_json::json!({"tool": "bash"}))
.transcript
.unwrap()
else {
panic!("approval row")
};
// A row after the one being rolled back, so removing the approval
// (index 1) shifts this one from index 2 down to index 1 — proving
// `row_index` gets reindexed, not just `rows`.
projection.apply(
2,
&event(serde_json::json!({
"sessionUpdate": "user_message_chunk",
"messageId": "m2",
"content": {"type": "text", "text": "hey"},
})),
);
assert_eq!(projection.snapshot(0).rows.len(), 3);
projection.rollback_row(&row.id);
assert_eq!(projection.snapshot(0).rows.len(), 2);
assert_eq!(projection.snapshot(0).rows[0].kind, TranscriptRowKind::User);
assert_eq!(projection.snapshot(0).rows[1].kind, TranscriptRowKind::User);
// The shifted row must still be independently patchable at its new
// index.
let update = projection.apply(
3,
&event(serde_json::json!({
"sessionUpdate": "user_message_chunk",
"messageId": "m2",
"content": {"type": "text", "text": " there"},
})),
);
let TranscriptDelta::Patch { row } = update.transcript.unwrap() else {
panic!("patched row")
};
assert_eq!(row.data.pointer("/content/text").unwrap(), "hey there");
assert_eq!(projection.snapshot(0).rows.len(), 2);
}
#[test]
fn rollback_row_is_a_no_op_for_an_unknown_id() {
let mut projection = Projection::default();
projection.apply(
1,
&event(serde_json::json!({
"sessionUpdate": "user_message_chunk",
"content": {"type": "text", "text": "hi"},
})),
);
projection.rollback_row("approval:does-not-exist");
assert_eq!(projection.snapshot(0).rows.len(), 1);
}
#[test]
fn recovered_approval_is_historical_not_pending() {
let mut live = Projection::default();
let TranscriptDelta::Append { row } = live
.request_approval("daemon-nonce".into(), serde_json::json!({"tool":"bash"}))
.transcript
.unwrap()
else {
panic!("approval row")
};
let mut recovered = Projection::default();
let update = recovered.restore_approval(row.clone());
assert_eq!(
update.transcript,
Some(TranscriptDelta::Append { row: row.clone() })
);
assert!(update.state.pending_approval.is_none());
assert_eq!(recovered.snapshot(0).rows, vec![row]);
let replay = serde_json::to_string(&recovered.snapshot(1).rows).unwrap();
assert!(!replay.contains("daemon-nonce"));
}
#[test]
fn stale_terminal_turn_cannot_clear_newer_approval() {
let mut projection = Projection::default();
let (first, _) = projection.begin_turn();
projection.turn_lifecycle(TurnLifecycle::Cancel, first);
let (second, _) = projection.begin_turn();
projection.request_approval("second-nonce".into(), serde_json::json!({"tool":"bash"}));
assert!(projection
.turn_lifecycle(TurnLifecycle::Finish, first)
.is_none());
assert_eq!(projection.active_turn(), Some(second));
assert_eq!(
projection.state().pending_approval.unwrap()["nonce"],
"second-nonce"
);
}
}
File diff suppressed because it is too large Load Diff
+372 -13
View File
@@ -7,8 +7,8 @@
//! advertised but with empty metadata (handled by the caller, not here).
//!
//! A structured session (#135) also owns `<id>.events.jsonl` next to its
//! sidecar: every `session/update` notification its adapter sent, one JSON
//! object per line, appended before it is broadcast. The log is the source of
//! sidecar: every `session/update` notification and daemon approval row is one
//! JSON object, appended before it is broadcast. The log is the source of
//! truth for the transcript; `seq` is the line's 1-based position, so a
//! client that holds seq `n` asks for [`EventCursor::Since`]`(n)`. A line
//! that does not parse — including the partial line a hard kill leaves at
@@ -21,7 +21,7 @@ use std::io::{BufRead, BufReader, Read, Seek, SeekFrom, Write};
use std::path::{Path, PathBuf};
use maestro_acp::schema::SessionNotification;
use maestro_protocol::Session;
use maestro_protocol::{Session, TranscriptRow};
use uuid::Uuid;
#[derive(Debug, thiserror::Error)]
@@ -112,7 +112,7 @@ impl SidecarStore {
}
file.seek(SeekFrom::End(0))?;
Ok(EventLog {
file,
file: Box::new(file),
last_seq: lines,
})
}
@@ -150,10 +150,10 @@ impl SidecarStore {
}
continue;
}
let Ok(notification) = serde_json::from_slice::<SessionNotification>(&line) else {
let Ok(event) = PersistedEvent::decode(&line) else {
continue;
};
out.push_back(LoggedEvent { seq, notification });
out.push_back(LoggedEvent { seq, event });
match cursor {
EventCursor::Since(_) if out.len() >= limit => break,
EventCursor::Before(_) if out.len() > limit => {
@@ -190,6 +190,7 @@ impl SidecarStore {
/// whichever does not exist.
pub fn delete(&self, id: Uuid) -> Result<(), SidecarError> {
remove_if_present(&self.events_path_for(id))?;
remove_if_present(&self.dir.join(format!("{id}.approvals.jsonl")))?;
remove_if_present(&self.recovery_path_for(id))?;
remove_if_present(&self.path_for(id))
}
@@ -251,18 +252,126 @@ pub enum EventCursor {
pub struct LoggedEvent {
/// 1-based line position in the log; the cursor for [`EventCursor`].
pub seq: u64,
pub notification: SessionNotification,
pub event: PersistedEvent,
}
/// A single entry in the structured event history.
///
/// ACP notifications retain their original JSON representation, so existing
/// event logs remain readable. Daemon-generated rows use an envelope so they
/// can share the same cursor sequence without pretending to be ACP traffic.
#[derive(Debug, Clone, PartialEq, serde::Serialize, serde::Deserialize)]
#[serde(tag = "maestroEvent", rename_all = "snake_case")]
pub enum PersistedEvent {
Approval {
row: TranscriptRow,
},
#[serde(skip)]
Notification(Box<SessionNotification>),
}
impl PersistedEvent {
fn decode(line: &[u8]) -> Result<Self, serde_json::Error> {
serde_json::from_slice::<SessionNotification>(line)
.map(|notification| Self::Notification(Box::new(notification)))
.or_else(|_| serde_json::from_slice(line))
}
}
impl From<SessionNotification> for PersistedEvent {
fn from(notification: SessionNotification) -> Self {
Self::Notification(Box::new(notification))
}
}
impl From<&SessionNotification> for PersistedEvent {
fn from(notification: &SessionNotification) -> Self {
Self::Notification(Box::new(notification.clone()))
}
}
impl From<&PersistedEvent> for PersistedEvent {
fn from(event: &PersistedEvent) -> Self {
event.clone()
}
}
/// What an [`EventLog`] appends onto: something seekable by length and
/// truncatable back to one, so a failed append can be rolled off the end
/// again. A real [`File`] in production; an in-memory fault double in tests
/// (#184) — the same [`EventLog::append`] rollback logic runs over both, so
/// the behavior under test is the production behavior, not a stand-in for
/// it.
pub(crate) trait LogSink: Write + Send {
fn len(&mut self) -> std::io::Result<u64>;
fn truncate(&mut self, len: u64) -> std::io::Result<()>;
}
impl LogSink for File {
fn len(&mut self) -> std::io::Result<u64> {
Ok(self.metadata()?.len())
}
fn truncate(&mut self, len: u64) -> std::io::Result<()> {
self.set_len(len)
}
}
/// Outcome of a failed [`EventLog::append`].
#[derive(Debug, thiserror::Error)]
pub enum AppendError {
/// The write or flush failed, but the log was rolled back to exactly the
/// byte it was at before the call: `last_seq` and the file agree, and
/// this same [`EventLog`] is safe to keep appending to. The event itself
/// was never persisted or broadcast.
#[error("event append failed, rolled back cleanly: {0}")]
Failed(#[source] SidecarError),
/// The append failed *and* the rollback did not restore the log to its
/// prior length, so the file's line count and `last_seq` may now
/// disagree. The caller must not append to this [`EventLog`] again —
/// treat the session as dead rather than promise consistency it cannot
/// keep.
#[error("event append failed ({cause}) and rollback did not restore consistency ({rollback})")]
Fatal {
cause: SidecarError,
rollback: SidecarError,
},
}
impl From<SidecarError> for AppendError {
fn from(error: SidecarError) -> Self {
AppendError::Failed(error)
}
}
/// An open, append-only handle on one session's event log. Obtained from
/// [`SidecarStore::open_events`]; held for the session's lifetime.
#[derive(Debug)]
pub struct EventLog {
file: File,
file: Box<dyn LogSink>,
last_seq: u64,
}
impl std::fmt::Debug for EventLog {
fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
f.debug_struct("EventLog")
.field("last_seq", &self.last_seq)
.finish_non_exhaustive()
}
}
impl EventLog {
/// Wrap an arbitrary sink as an [`EventLog`], for tests that need to
/// inject a deterministic append failure (#184) — a partial write, a
/// flush that fails after the bytes are already down, or a rollback that
/// cannot itself complete.
#[cfg(test)]
pub(crate) fn for_test(file: impl LogSink + 'static, last_seq: u64) -> Self {
Self {
file: Box::new(file),
last_seq,
}
}
/// The seq of the last line, `0` for an empty log.
pub fn last_seq(&self) -> u64 {
self.last_seq
@@ -272,12 +381,262 @@ impl EventLog {
/// handed to the kernel in a single write before this returns, so a
/// caller that broadcasts afterwards never announces an event the log
/// would not replay — short of the machine losing power.
pub fn append(&mut self, notification: &SessionNotification) -> Result<u64, SidecarError> {
let mut line = serde_json::to_vec(notification)?;
///
/// A `write_all` that errors partway can leave an unterminated prefix on
/// disk; a `flush` that errors after a *complete* `write_all` leaves
/// ambiguity about whether that full, well-formed line should count —
/// counting it would silently shift every later seq by one, since
/// `last_seq` was never bumped for it. Both cases are rolled back to the
/// exact pre-append length before returning, so `last_seq` and the file
/// never disagree and the same [`EventLog`] can be appended to again
/// immediately. If the rollback itself fails, the log's consistency can
/// no longer be promised and [`AppendError::Fatal`] is returned instead.
pub fn append(&mut self, event: impl Into<PersistedEvent>) -> Result<u64, AppendError> {
let event = event.into();
let mut line = match &event {
PersistedEvent::Notification(notification) => serde_json::to_vec(notification)
.map_err(|e| AppendError::from(SidecarError::from(e)))?,
_ => {
serde_json::to_vec(&event).map_err(|e| AppendError::from(SidecarError::from(e)))?
}
};
line.push(b'\n');
self.file.write_all(&line)?;
self.file.flush()?;
let pre_len = self
.file
.len()
.map_err(|e| AppendError::from(SidecarError::from(e)))?;
if let Err(cause) = self.file.write_all(&line) {
return Err(self.rollback_after(pre_len, cause));
}
if let Err(cause) = self.file.flush() {
return Err(self.rollback_after(pre_len, cause));
}
self.last_seq += 1;
Ok(self.last_seq)
}
fn rollback_after(&mut self, pre_len: u64, cause: std::io::Error) -> AppendError {
match self.file.truncate(pre_len) {
Ok(()) => AppendError::Failed(SidecarError::Io(cause)),
Err(rollback) => AppendError::Fatal {
cause: SidecarError::Io(cause),
rollback: SidecarError::Io(rollback),
},
}
}
}
/// Bytes shared between a [`FaultyLogSink`] moved into an [`EventLog`] and
/// the test that planted it, so the test can inspect what actually ended up
/// on the "disk" after a fault without reaching into `EventLog`'s private
/// `file` field.
#[cfg(test)]
#[derive(Clone, Default)]
pub(crate) struct SharedBuf(std::sync::Arc<std::sync::Mutex<Vec<u8>>>);
#[cfg(test)]
impl SharedBuf {
pub(crate) fn snapshot(&self) -> Vec<u8> {
self.0.lock().unwrap().clone()
}
}
/// An in-memory [`LogSink`] double that injects one deterministic fault — a
/// short write, a flush that fails after a full write, or a rollback that
/// cannot itself complete — then behaves like a healthy sink from then on.
/// Lets a test prove the *same* [`EventLog`] recovers, rather than one with a
/// handle swapped in from a healthy disk, which production never does
/// (#184).
#[cfg(test)]
#[derive(Default)]
pub(crate) struct FaultyLogSink {
pub(crate) buf: SharedBuf,
/// Accept this many bytes on the next `write` call, then fail the
/// *following* call — a short write followed by an error, exactly what
/// `write_all` sees from a real fd that dies mid-line. `Write::write`
/// must not report an error on a call that made partial progress, so
/// the fault is delivered one call later, via `pending_write_error`.
pub(crate) fail_write_after: Option<usize>,
pub(crate) pending_write_error: bool,
pub(crate) fail_flush_once: bool,
pub(crate) fail_truncate: bool,
}
#[cfg(test)]
impl Write for FaultyLogSink {
fn write(&mut self, data: &[u8]) -> std::io::Result<usize> {
if self.pending_write_error {
self.pending_write_error = false;
return Err(std::io::Error::other("fault: write failed mid-line"));
}
if let Some(allowed) = self.fail_write_after.take() {
let allowed = allowed.min(data.len());
if allowed == 0 {
// No progress was made, so it's fine to report the error
// from this same call.
return Err(std::io::Error::other("fault: write failed mid-line"));
}
self.buf
.0
.lock()
.unwrap()
.extend_from_slice(&data[..allowed]);
self.pending_write_error = true;
return Ok(allowed);
}
self.buf.0.lock().unwrap().extend_from_slice(data);
Ok(data.len())
}
fn flush(&mut self) -> std::io::Result<()> {
if self.fail_flush_once {
self.fail_flush_once = false;
return Err(std::io::Error::other(
"fault: flush failed after a full write",
));
}
Ok(())
}
}
#[cfg(test)]
impl LogSink for FaultyLogSink {
fn len(&mut self) -> std::io::Result<u64> {
Ok(self.buf.0.lock().unwrap().len() as u64)
}
fn truncate(&mut self, len: u64) -> std::io::Result<()> {
if self.fail_truncate {
return Err(std::io::Error::other("fault: rollback truncate failed"));
}
self.buf.0.lock().unwrap().truncate(len as usize);
Ok(())
}
}
#[cfg(test)]
mod tests {
use super::*;
fn sample_event(id: &str) -> PersistedEvent {
PersistedEvent::Approval {
row: TranscriptRow {
id: id.to_owned(),
kind: maestro_protocol::TranscriptRowKind::Approval,
data: serde_json::json!({"tool": "bash"}),
},
}
}
/// Replays `bytes` as a real on-disk log via the production read path,
/// so "live and durable agree" is checked with the same code a
/// reconnecting client would hit, not a hand-rolled line splitter.
fn replay(bytes: &[u8]) -> Vec<LoggedEvent> {
let dir = tempfile::tempdir().unwrap();
let store = SidecarStore::new(dir.path());
let id = Uuid::new_v4();
std::fs::create_dir_all(store.dir()).unwrap();
std::fs::write(store.events_path_for(id), bytes).unwrap();
store.read_events(id, EventCursor::Since(0), 100).unwrap()
}
#[test]
fn append_rolls_back_a_partial_write_and_the_same_instance_recovers() {
let buf = SharedBuf::default();
let writer = FaultyLogSink {
buf: buf.clone(),
fail_write_after: Some(5),
..Default::default()
};
let mut log = EventLog::for_test(writer, 0);
let err = log.append(sample_event("r1")).unwrap_err();
assert!(
matches!(err, AppendError::Failed(_)),
"a repairable rollback must not be reported fatal"
);
assert_eq!(log.last_seq(), 0, "cursor must stay put on a failed append");
assert!(
buf.snapshot().is_empty(),
"the partial prefix a failed write_all leaves behind must be truncated away"
);
let seq = log
.append(sample_event("r1"))
.expect("the same instance must be able to append again after a clean rollback");
assert_eq!(seq, 1);
assert_eq!(log.last_seq(), 1);
let bytes = buf.snapshot();
assert_eq!(
bytes.iter().filter(|&&b| b == b'\n').count(),
1,
"exactly one well-formed line, not a partial prefix glued to the recovered append"
);
let replayed = replay(&bytes);
assert_eq!(replayed.len(), 1);
assert_eq!(replayed[0].seq, log.last_seq());
let PersistedEvent::Approval { row } = &replayed[0].event else {
panic!("approval event");
};
assert_eq!(row.id, "r1");
}
#[test]
fn append_rolls_back_when_flush_fails_after_a_full_write() {
let buf = SharedBuf::default();
let writer = FaultyLogSink {
buf: buf.clone(),
fail_flush_once: true,
..Default::default()
};
let mut log = EventLog::for_test(writer, 0);
let err = log.append(sample_event("r1")).unwrap_err();
assert!(matches!(err, AppendError::Failed(_)));
assert_eq!(log.last_seq(), 0);
assert!(
buf.snapshot().is_empty(),
"a full-but-unflushed line must not survive uncounted by last_seq — it would \
silently shift every later seq by one"
);
let seq = log
.append(sample_event("r1"))
.expect("the same instance must recover once flush stops failing");
assert_eq!(seq, 1);
let bytes = buf.snapshot();
let replayed = replay(&bytes);
assert_eq!(replayed.len(), 1);
assert_eq!(replayed[0].seq, log.last_seq());
}
#[test]
fn append_reports_fatal_when_rollback_cannot_restore_consistency() {
let buf = SharedBuf::default();
let writer = FaultyLogSink {
buf,
fail_write_after: Some(3),
fail_truncate: true,
..Default::default()
};
let mut log = EventLog::for_test(writer, 0);
let err = log.append(sample_event("r1")).unwrap_err();
assert!(
matches!(err, AppendError::Fatal { .. }),
"an unrepairable rollback must be surfaced as fatal, never silently treated as \
a clean recoverable failure"
);
assert_eq!(
log.last_seq(),
0,
"a fatal append must not advance the cursor either"
);
}
}
+245 -5
View File
@@ -83,7 +83,10 @@ impl Fake {
}
fn event_text(event: &maestro_daemon::sidecar::LoggedEvent) -> &str {
match &event.notification.update {
let maestro_daemon::sidecar::PersistedEvent::Notification(notification) = &event.event else {
panic!("unexpected daemon event")
};
match &notification.update {
SessionUpdate::AgentMessageChunk(chunk) => match &chunk.content {
ContentBlock::Text(text) => &text.text,
other => panic!("unexpected content {other:?}"),
@@ -133,6 +136,7 @@ async fn structured_spawn_starts_the_adapter_in_cwd_with_session_env() {
assert_eq!(session.view, SessionView::Structured);
assert_eq!(session.status, SessionStatus::Running);
assert_eq!(session.activity, ActivityState::Idle);
let acp_id = session.acp_session_id.clone().expect("acp session id");
assert!(acp_id.starts_with("fake-"), "{acp_id}");
@@ -201,6 +205,77 @@ async fn structured_spawn_starts_the_adapter_in_cwd_with_session_env() {
.exists());
}
#[tokio::test]
async fn structured_activity_tracks_turn_completion_and_failure() {
let fx = DaemonFixture::new();
let fake = Fake::new();
let cwd = TempDir::new().unwrap();
let mut env = fake.env();
env.insert("FAKE_ACP_DELAY_PROMPT_MS".into(), "100".into());
env.insert("FAKE_ACP_PROMPT_FAIL_ONCE".into(), "1".into());
let session = fx
.manager
.create(structured(cwd.path(), env))
.await
.unwrap();
assert!(matches!(
fx.manager
.queue_acp_prompt(session.id, "fail".into())
.await
.unwrap(),
maestro_daemon::session::QueueAcpPrompt::Queued
));
assert_eq!(
fx.manager
.snapshot()
.await
.into_iter()
.find(|candidate| candidate.id == session.id)
.unwrap()
.activity,
ActivityState::Active
);
wait_for_async(Duration::from_secs(2), || async {
fx.manager
.snapshot()
.await
.into_iter()
.find(|candidate| candidate.id == session.id)
.is_some_and(|candidate| candidate.activity == ActivityState::Error)
})
.await;
assert!(matches!(
fx.manager
.queue_acp_prompt(session.id, "succeed".into())
.await
.unwrap(),
maestro_daemon::session::QueueAcpPrompt::Queued
));
assert_eq!(
fx.manager
.snapshot()
.await
.into_iter()
.find(|candidate| candidate.id == session.id)
.unwrap()
.activity,
ActivityState::Active
);
wait_for_async(Duration::from_secs(2), || async {
fx.manager
.snapshot()
.await
.into_iter()
.find(|candidate| candidate.id == session.id)
.is_some_and(|candidate| candidate.activity == ActivityState::Idle)
})
.await;
fx.manager.kill(session.id, false).await.unwrap();
}
#[tokio::test]
async fn projection_snapshot_and_live_stream_share_one_fold() {
let fx = DaemonFixture::new();
@@ -232,11 +307,14 @@ async fn projection_snapshot_and_live_stream_share_one_fold() {
assert_eq!(snapshot.rows[0].kind, TranscriptRowKind::AgentText);
assert_eq!(state.turn, Default::default());
assert!(fx
let maestro_daemon::session::BeginAcpTurn::Started(_, turn) = fx
.manager
.apply_acp_turn_lifecycle(session.id, TurnLifecycle::Begin)
.begin_acp_turn(session.id)
.await
.expect("begin lifecycle"));
.expect("begin lifecycle")
else {
panic!("structured turn should start");
};
let begin = tokio::time::timeout(Duration::from_secs(5), live.recv())
.await
.expect("begin lifecycle broadcast")
@@ -264,7 +342,7 @@ async fn projection_snapshot_and_live_stream_share_one_fold() {
assert!(fx
.manager
.apply_acp_turn_lifecycle(session.id, TurnLifecycle::Finish)
.apply_acp_turn_lifecycle(session.id, TurnLifecycle::Finish, turn)
.await
.expect("finish lifecycle"));
let finish = tokio::time::timeout(Duration::from_secs(5), live.recv())
@@ -277,6 +355,148 @@ async fn projection_snapshot_and_live_stream_share_one_fold() {
fx.manager.kill(session.id, false).await.expect("kill");
}
#[tokio::test]
async fn stale_turn_completion_cannot_end_the_next_turn() {
let fx = DaemonFixture::new();
let fake = Fake::new();
let cwd = TempDir::new().unwrap();
let session = fx
.manager
.create(structured(cwd.path(), fake.env()))
.await
.unwrap();
let maestro_daemon::session::BeginAcpTurn::Started(_, first) =
fx.manager.begin_acp_turn(session.id).await.unwrap()
else {
panic!("first turn should start");
};
assert!(fx.manager.cancel_acp_turn(session.id).await.unwrap());
let maestro_daemon::session::BeginAcpTurn::Started(_, second) =
fx.manager.begin_acp_turn(session.id).await.unwrap()
else {
panic!("second turn should start");
};
assert!(second > first);
assert!(!fx
.manager
.apply_acp_turn_lifecycle(session.id, TurnLifecycle::Finish, first)
.await
.unwrap());
assert!(matches!(
fx.manager.begin_acp_turn(session.id).await.unwrap(),
maestro_daemon::session::BeginAcpTurn::InFlight
));
assert!(fx
.manager
.apply_acp_turn_lifecycle(session.id, TurnLifecycle::Finish, second)
.await
.unwrap());
fx.manager.kill(session.id, false).await.unwrap();
}
#[tokio::test]
async fn cancelled_prompt_that_never_settles_fails_the_session() {
let fx = DaemonFixture::new();
let fake = Fake::new();
let cwd = TempDir::new().unwrap();
let mut env = fake.env();
env.insert("FAKE_ACP_PERMISSION_AFTER_CANCEL".into(), "1".into());
let session = fx
.manager
.create(structured(cwd.path(), env))
.await
.unwrap();
assert!(matches!(
fx.manager
.queue_acp_prompt(session.id, "first".into())
.await
.unwrap(),
maestro_daemon::session::QueueAcpPrompt::Queued
));
assert!(fx.manager.cancel_acp_turn(session.id).await.unwrap());
let second = fx
.manager
.queue_acp_prompt(session.id, "second".into())
.await
.unwrap();
assert!(
matches!(second, maestro_daemon::session::QueueAcpPrompt::Unavailable),
"{second:?}"
);
assert_eq!(
fx.manager.status(session.id).await,
Some(SessionStatus::Dead)
);
assert!(!fake.log().contains("permission response"));
}
#[tokio::test]
async fn cancellation_settlement_allows_the_next_prompt_permission() {
let fx = DaemonFixture::new();
let fake = Fake::new();
let cwd = TempDir::new().unwrap();
let mut env = fake.env();
env.insert("FAKE_ACP_REQUEST_PERMISSION".into(), "1".into());
env.insert("FAKE_ACP_CANCEL_SETTLES".into(), "1".into());
let session = fx
.manager
.create(structured(cwd.path(), env))
.await
.unwrap();
assert!(matches!(
fx.manager
.queue_acp_prompt(session.id, "first".into())
.await
.unwrap(),
maestro_daemon::session::QueueAcpPrompt::Queued
));
assert!(fx.manager.cancel_acp_turn(session.id).await.unwrap());
assert!(matches!(
fx.manager
.queue_acp_prompt(session.id, "second".into())
.await
.unwrap(),
maestro_daemon::session::QueueAcpPrompt::Queued
));
let nonce = tokio::time::timeout(Duration::from_secs(2), async {
loop {
let (_, state, _) = fx
.manager
.subscribe_acp_projection(session.id, EventCursor::Since(0))
.await
.unwrap()
.unwrap();
if let Some(pending) = state.pending_approval {
break pending["nonce"].as_str().unwrap().to_owned();
}
tokio::time::sleep(Duration::from_millis(10)).await;
}
})
.await
.unwrap();
assert!(fx
.manager
.resolve_acp_approval(
session.id,
&nonce,
maestro_daemon::session::ApprovalDecision::AllowOnce,
)
.await
.unwrap()
.is_some());
wait_for_async(Duration::from_secs(2), || {
let log = fake.log();
async move { log.contains("permission response") }
})
.await;
fx.manager.kill(session.id, false).await.unwrap();
}
#[tokio::test]
async fn recovery_loads_the_acp_session_without_duplicating_history() {
let fx = DaemonFixture::new();
@@ -328,6 +548,16 @@ async fn recovery_loads_the_acp_session_without_duplicating_history() {
revived.status(session.id).await,
Some(SessionStatus::Running)
);
assert_eq!(
revived
.snapshot()
.await
.into_iter()
.find(|candidate| candidate.id == session.id)
.unwrap()
.activity,
ActivityState::Idle
);
let log = fake.log();
assert!(
log.contains(&format!("recv session/load sessionId={acp_id}")),
@@ -777,6 +1007,16 @@ async fn adapter_exit_marks_the_session_dead_and_keeps_it_listed() {
fx.manager.status(session.id).await == Some(SessionStatus::Dead)
})
.await;
assert_eq!(
fx.manager
.snapshot()
.await
.into_iter()
.find(|candidate| candidate.id == session.id)
.unwrap()
.activity,
ActivityState::Error
);
assert!(fx.manager.has(session.id).await);
assert!(
!fx.sidecar_dir()
+656 -1
View File
@@ -4,12 +4,17 @@ mod common;
use std::collections::HashMap;
use std::net::{Ipv4Addr, SocketAddr};
use std::path::Path;
use std::time::Duration;
use common::DaemonFixture;
use futures_util::{SinkExt, StreamExt};
use maestro_daemon::local_api;
use maestro_protocol::{ApiError, CreateSessionRequest, Session, WorktreeSpec};
use maestro_protocol::{
ActivityState, ApiError, CreateSessionRequest, Session, SessionStatus, SessionView,
WorktreeSpec,
};
use tempfile::TempDir;
use tokio::net::TcpListener;
use tokio::sync::oneshot;
use tokio_tungstenite::tungstenite::Message as WsMessage;
@@ -86,6 +91,589 @@ fn ws_url(addr: SocketAddr, path: &str) -> String {
format!("ws://{addr}{path}")
}
struct FakeAcp {
dir: TempDir,
}
impl FakeAcp {
fn new() -> Self {
use std::os::unix::fs::PermissionsExt;
let dir = TempDir::new().unwrap();
let bin = dir.path().join("bin");
std::fs::create_dir_all(&bin).unwrap();
let adapter = bin.join("claude-code-acp");
std::fs::copy(
Path::new(env!("CARGO_MANIFEST_DIR")).join("../maestro-acp/tests/fake_acp_agent.py"),
&adapter,
)
.unwrap();
std::fs::set_permissions(&adapter, std::fs::Permissions::from_mode(0o755)).unwrap();
Self { dir }
}
fn env(&self) -> HashMap<String, String> {
HashMap::from([
(
"PATH".into(),
format!(
"{}:{}",
self.dir.path().join("bin").display(),
std::env::var("PATH").unwrap()
),
),
(
"FAKE_ACP_LOG".into(),
self.dir.path().join("agent.log").display().to_string(),
),
])
}
fn pid(&self) -> u32 {
std::fs::read_to_string(self.dir.path().join("agent.log"))
.unwrap_or_default()
.lines()
.filter_map(|line| {
line.strip_prefix("start pid=")?
.split(' ')
.next()?
.parse()
.ok()
})
.next_back()
.expect("fake ACP started")
}
fn log(&self) -> String {
std::fs::read_to_string(self.dir.path().join("agent.log")).unwrap_or_default()
}
}
async fn structured_session(h: &Harness, fake: &FakeAcp) -> Uuid {
structured_session_with_env(h, fake.env()).await
}
async fn structured_session_with_env(h: &Harness, env: HashMap<String, String>) -> Uuid {
h.manager
.create(CreateSessionRequest {
view: SessionView::Structured,
cwd: "/tmp".into(),
command: "claude".into(),
env,
metadata: HashMap::new(),
worktree: None,
})
.await
.unwrap()
.id
}
async fn wait_for_session_activity(h: &Harness, id: Uuid, expected: ActivityState) {
tokio::time::timeout(Duration::from_secs(2), async {
loop {
let sessions: Vec<Session> = reqwest::get(url(h.addr, "/api/sessions"))
.await
.unwrap()
.json()
.await
.unwrap();
if sessions
.iter()
.find(|session| session.id == id)
.is_some_and(|session| session.activity == expected)
{
return;
}
tokio::time::sleep(Duration::from_millis(10)).await;
}
})
.await
.unwrap_or_else(|_| panic!("session {id} did not reach {expected:?}"));
}
#[tokio::test]
async fn structured_ws_sends_snapshot_then_live_projection() {
let h = spawn("naps-laptop").await;
let fake = FakeAcp::new();
let id = structured_session(&h, &fake).await;
let (mut ws, _) = connect_async(ws_url(h.addr, &format!("/api/structured/{id}")))
.await
.unwrap();
let first = ws.next().await.unwrap().unwrap().into_text().unwrap();
let second = ws.next().await.unwrap().unwrap().into_text().unwrap();
assert_eq!(
serde_json::from_str::<serde_json::Value>(&first).unwrap()["type"],
"transcript_snapshot"
);
assert_eq!(
serde_json::from_str::<serde_json::Value>(&second).unwrap()["type"],
"reduced_state"
);
let response = reqwest::Client::new()
.post(url(
h.addr,
&format!("/api/sessions/{id}/structured/prompt"),
))
.json(&serde_json::json!({"prompt": "hello"}))
.send()
.await
.unwrap();
assert_eq!(response.status(), reqwest::StatusCode::NO_CONTENT);
let live = tokio::time::timeout(Duration::from_secs(1), ws.next())
.await
.expect("projection update arrives on the subscribed stream")
.unwrap()
.unwrap()
.into_text()
.unwrap();
assert_eq!(
serde_json::from_str::<serde_json::Value>(&live).unwrap()["type"],
"projection_update"
);
}
#[tokio::test]
async fn structured_prompt_rejects_a_second_in_flight_turn_and_cancel_clears_it() {
let h = spawn("naps-laptop").await;
let fake = FakeAcp::new();
let mut env = fake.env();
env.insert("FAKE_ACP_DELAY_PROMPT_MS".into(), "200".into());
let id = structured_session_with_env(&h, env).await;
let client = reqwest::Client::new();
let first = client
.post(url(
h.addr,
&format!("/api/sessions/{id}/structured/prompt"),
))
.json(&serde_json::json!({"prompt": "first"}))
.send()
.await
.unwrap();
assert_eq!(first.status(), reqwest::StatusCode::NO_CONTENT);
let second = client
.post(url(
h.addr,
&format!("/api/sessions/{id}/structured/prompt"),
))
.json(&serde_json::json!({"prompt": "second"}))
.send()
.await
.unwrap();
assert_eq!(second.status(), reqwest::StatusCode::CONFLICT);
let error: ApiError = second.json().await.unwrap();
assert_eq!(error.error.code, "turn_in_flight");
let cancel = client
.post(url(
h.addr,
&format!("/api/sessions/{id}/structured/cancel"),
))
.send()
.await
.unwrap();
assert_eq!(cancel.status(), reqwest::StatusCode::NO_CONTENT);
let next = client
.post(url(
h.addr,
&format!("/api/sessions/{id}/structured/prompt"),
))
.json(&serde_json::json!({"prompt": "next"}))
.send()
.await
.unwrap();
assert_eq!(next.status(), reqwest::StatusCode::NO_CONTENT);
tokio::time::timeout(Duration::from_secs(2), async {
loop {
if fake.log().matches("recv session/prompt").count() == 2 {
break;
}
tokio::time::sleep(Duration::from_millis(10)).await;
}
})
.await
.expect("both prompt commands reach ACP");
let log = fake.log();
let first_prompt = log.find("recv session/prompt").unwrap();
let cancel = log.find("recv session/cancel").unwrap();
let second_prompt = log.rfind("recv session/prompt").unwrap();
assert!(first_prompt < cancel && cancel < second_prompt, "{log}");
}
#[tokio::test]
async fn structured_approval_is_secret_single_use_and_race_safe() {
let h = spawn("naps-laptop").await;
let fake = FakeAcp::new();
let mut env = fake.env();
env.insert("FAKE_ACP_REQUEST_PERMISSION".into(), "1".into());
let id = structured_session_with_env(&h, env).await;
let client = reqwest::Client::new();
let (mut ws, _) = connect_async(ws_url(h.addr, &format!("/api/structured/{id}")))
.await
.unwrap();
ws.next().await.unwrap().unwrap();
ws.next().await.unwrap().unwrap();
assert_eq!(
client
.post(url(
h.addr,
&format!("/api/sessions/{id}/structured/prompt")
))
.json(&serde_json::json!({"prompt":"needs permission"}))
.send()
.await
.unwrap()
.status(),
reqwest::StatusCode::NO_CONTENT
);
let nonce = tokio::time::timeout(Duration::from_secs(2), async {
loop {
let event: serde_json::Value =
serde_json::from_str(&ws.next().await.unwrap().unwrap().into_text().unwrap())
.unwrap();
if let Some(nonce) = event["state"]["pending_approval"]["nonce"].as_str() {
assert_eq!(event["transcript"]["row"]["kind"], "approval");
assert!(event["transcript"]["row"]["data"].get("nonce").is_none());
break nonce.to_owned();
}
}
})
.await
.unwrap();
wait_for_session_activity(&h, id, ActivityState::AwaitingInput).await;
assert!(
!fake.log().contains(&nonce),
"daemon nonce leaked to ACP agent"
);
let path = format!("/api/sessions/{id}/structured/approvals/{nonce}");
let (a, b) = tokio::join!(
client
.post(url(h.addr, &path))
.json(&serde_json::json!("allow_once"))
.send(),
client
.post(url(h.addr, &path))
.json(&serde_json::json!("deny"))
.send(),
);
assert_eq!(a.unwrap().status(), reqwest::StatusCode::NO_CONTENT);
assert_eq!(b.unwrap().status(), reqwest::StatusCode::NO_CONTENT);
assert_eq!(
client
.post(url(h.addr, &path))
.json(&serde_json::json!("deny"))
.send()
.await
.unwrap()
.status(),
reqwest::StatusCode::NO_CONTENT
);
tokio::time::timeout(Duration::from_secs(2), async {
loop {
if fake.log().contains("permission response") {
break;
}
tokio::time::sleep(Duration::from_millis(10)).await;
}
})
.await
.unwrap();
let log = fake.log();
assert!(
log.contains("\"optionId\": \"once\"") || log.contains("\"optionId\": \"deny\""),
"{log}"
);
tokio::time::timeout(Duration::from_secs(1), async {
loop {
let event: serde_json::Value =
serde_json::from_str(&ws.next().await.unwrap().unwrap().into_text().unwrap())
.unwrap();
if event["state"]["pending_approval"].is_null() {
break;
}
}
})
.await
.expect("resolved approval clears the reduced state");
wait_for_session_activity(&h, id, ActivityState::Idle).await;
assert_eq!(
client
.post(url(
h.addr,
&format!(
"/api/sessions/{}/structured/approvals/{nonce}",
Uuid::new_v4()
)
))
.json(&serde_json::json!("deny"))
.send()
.await
.unwrap()
.status(),
reqwest::StatusCode::NOT_FOUND
);
}
#[tokio::test]
async fn structured_prompt_failure_clears_the_turn() {
let h = spawn("naps-laptop").await;
let fake = FakeAcp::new();
let mut env = fake.env();
env.insert("FAKE_ACP_PROMPT_FAIL_ONCE".into(), "1".into());
let id = structured_session_with_env(&h, env).await;
let client = reqwest::Client::new();
let path = format!("/api/sessions/{id}/structured/prompt");
let (mut ws, _) = connect_async(ws_url(h.addr, &format!("/api/structured/{id}")))
.await
.unwrap();
ws.next().await.unwrap().unwrap();
ws.next().await.unwrap().unwrap();
let first = client
.post(url(h.addr, &path))
.json(&serde_json::json!({"prompt":"fails"}))
.send()
.await
.unwrap();
assert_eq!(first.status(), reqwest::StatusCode::NO_CONTENT);
let mut saw_active_turn = false;
tokio::time::timeout(Duration::from_secs(1), async {
while let Some(Ok(message)) = ws.next().await {
let event: serde_json::Value =
serde_json::from_str(&message.into_text().unwrap()).unwrap();
let turn = &event["state"]["turn"];
if turn["responding"] == true {
saw_active_turn = true;
}
if saw_active_turn && turn["responding"] == false && turn["thinking"] == false {
return;
}
}
panic!("structured websocket closed before the failed turn cleared");
})
.await
.expect("failed turn clears promptly");
wait_for_session_activity(&h, id, ActivityState::Error).await;
let second = client
.post(url(h.addr, &path))
.json(&serde_json::json!({"prompt":"available"}))
.send()
.await
.unwrap();
assert_eq!(second.status(), reqwest::StatusCode::NO_CONTENT);
assert!(second.bytes().await.unwrap().is_empty());
}
#[tokio::test]
async fn structured_prompt_rejects_unknown_and_unavailable_sessions() {
let h = spawn("naps-laptop").await;
let client = reqwest::Client::new();
let prompt = |id| {
client
.post(url(
h.addr,
&format!("/api/sessions/{id}/structured/prompt"),
))
.json(&serde_json::json!({"prompt": "hello"}))
};
let unknown = prompt(Uuid::new_v4()).send().await.unwrap();
assert_eq!(unknown.status(), reqwest::StatusCode::NOT_FOUND);
let error: ApiError = unknown.json().await.unwrap();
assert_eq!(error.error.code, "not_found");
let terminal = h
.manager
.create(CreateSessionRequest {
view: SessionView::Terminal,
cwd: "/tmp".into(),
command: "sh".into(),
env: HashMap::new(),
metadata: HashMap::new(),
worktree: None,
})
.await
.unwrap();
let non_structured = prompt(terminal.id).send().await.unwrap();
assert_eq!(non_structured.status(), reqwest::StatusCode::CONFLICT);
let error: ApiError = non_structured.json().await.unwrap();
assert_eq!(error.error.code, "not_structured");
let fake = FakeAcp::new();
let dead = structured_session(&h, &fake).await;
h.manager.shutdown().await;
assert_eq!(h.manager.status(dead).await, Some(SessionStatus::Dead));
let unavailable = prompt(dead).send().await.unwrap();
assert_eq!(unavailable.status(), reqwest::StatusCode::CONFLICT);
let error: ApiError = unavailable.json().await.unwrap();
assert_eq!(error.error.code, "not_structured");
}
#[tokio::test]
async fn structured_replay_honors_cursors_and_limit() {
let h = spawn("naps-laptop").await;
let fake = FakeAcp::new();
let id = structured_session(&h, &fake).await;
let client = reqwest::Client::new();
let events: Vec<serde_json::Value> = tokio::time::timeout(Duration::from_secs(2), async {
loop {
let events: Vec<serde_json::Value> = client
.get(url(
h.addr,
&format!("/api/sessions/{id}/structured/replay?since=0&limit=1"),
))
.send()
.await
.unwrap()
.json()
.await
.unwrap();
if !events.is_empty() {
return events;
}
tokio::time::sleep(Duration::from_millis(10)).await;
}
})
.await
.unwrap();
assert_eq!(events.len(), 1);
assert_eq!(events[0]["cursor"], 1);
assert!(events[0].get("notification").is_none());
let invalid = client
.get(url(
h.addr,
&format!("/api/sessions/{id}/structured/replay?since=0&before=2"),
))
.send()
.await
.unwrap();
assert_eq!(invalid.status(), reqwest::StatusCode::BAD_REQUEST);
}
#[tokio::test]
async fn structured_replay_pages_projected_events_in_cursor_order() {
let h = spawn("naps-laptop").await;
let fake = FakeAcp::new();
let id = structured_session(&h, &fake).await;
let client = reqwest::Client::new();
unsafe { libc::kill(fake.pid() as libc::pid_t, libc::SIGUSR1) };
let events: Vec<serde_json::Value> = tokio::time::timeout(Duration::from_secs(2), async {
loop {
let events: Vec<serde_json::Value> = client
.get(url(
h.addr,
&format!("/api/sessions/{id}/structured/replay?before=99&limit=2"),
))
.send()
.await
.unwrap()
.json()
.await
.unwrap();
if events.len() == 2 {
return events;
}
tokio::time::sleep(Duration::from_millis(10)).await;
}
})
.await
.unwrap();
assert_eq!(
events
.iter()
.map(|event| event["cursor"].as_u64().unwrap())
.collect::<Vec<_>>(),
vec![1, 2]
);
let boundary: Vec<serde_json::Value> = client
.get(url(
h.addr,
&format!("/api/sessions/{id}/structured/replay?before=2&limit=10"),
))
.send()
.await
.unwrap()
.json()
.await
.unwrap();
assert_eq!(boundary.len(), 1);
assert_eq!(boundary[0]["cursor"], 1);
}
#[tokio::test]
async fn structured_replay_rejects_unknown_terminal_and_dead_sessions() {
let h = spawn("naps-laptop").await;
let client = reqwest::Client::new();
let unknown = client
.get(url(
h.addr,
&format!("/api/sessions/{}/structured/replay", Uuid::new_v4()),
))
.send()
.await
.unwrap();
assert_eq!(unknown.status(), reqwest::StatusCode::NOT_FOUND);
let terminal = h
.manager
.create(CreateSessionRequest {
view: SessionView::Terminal,
cwd: "/tmp".into(),
command: "sh".into(),
env: HashMap::new(),
metadata: HashMap::new(),
worktree: None,
})
.await
.unwrap();
let response = client
.get(url(
h.addr,
&format!("/api/sessions/{}/structured/replay", terminal.id),
))
.send()
.await
.unwrap();
assert_eq!(response.status(), reqwest::StatusCode::CONFLICT);
let fake = FakeAcp::new();
let dead = structured_session(&h, &fake).await;
h.manager.shutdown().await;
assert_eq!(h.manager.status(dead).await, Some(SessionStatus::Dead));
let response = client
.get(url(
h.addr,
&format!("/api/sessions/{dead}/structured/replay"),
))
.send()
.await
.unwrap();
assert_eq!(response.status(), reqwest::StatusCode::CONFLICT);
let error: ApiError = response.json().await.unwrap();
assert_eq!(error.error.code, "not_structured");
}
#[tokio::test]
async fn structured_ws_rejects_a_dead_session() {
let h = spawn("naps-laptop").await;
let fake = FakeAcp::new();
let id = structured_session(&h, &fake).await;
h.manager.shutdown().await;
assert_eq!(h.manager.status(id).await, Some(SessionStatus::Dead));
let err = connect_async(ws_url(h.addr, &format!("/api/structured/{id}")))
.await
.expect_err("dead structured session must not upgrade");
let message = err.to_string();
assert!(
message.contains("409") || message.to_lowercase().contains("conflict"),
"{message}"
);
}
#[tokio::test]
async fn info_returns_daemon_identity() {
let h = spawn("naps-laptop").await;
@@ -789,6 +1377,73 @@ async fn events_ws_sends_a_snapshot_then_streams_deltas() {
}
}
#[tokio::test]
async fn events_ws_streams_structured_activity_without_hook_overrides() {
let h = spawn("konishi").await;
let fake = FakeAcp::new();
let mut env = fake.env();
env.insert("FAKE_ACP_DELAY_PROMPT_MS".into(), "500".into());
let (mut ws, _) = connect_async(ws_url(h.addr, "/api/events"))
.await
.expect("events ws");
assert!(matches!(
next_event(&mut ws).await,
maestro_protocol::CatalogEvent::Snapshot { .. }
));
let id = structured_session_with_env(&h, env).await;
let maestro_protocol::CatalogEvent::SessionAdded { session } = next_event(&mut ws).await else {
panic!("expected structured SessionAdded")
};
assert_eq!(session.activity, ActivityState::Idle);
let client = reqwest::Client::new();
let prompt = client
.post(url(
h.addr,
&format!("/api/sessions/{id}/structured/prompt"),
))
.json(&serde_json::json!({"prompt": "hello"}))
.send()
.await
.unwrap();
assert_eq!(prompt.status(), reqwest::StatusCode::NO_CONTENT);
let report = client
.post(url(h.addr, &format!("/api/sessions/{id}/activity")))
.json(&serde_json::json!({"state": "Error"}))
.send()
.await
.unwrap();
assert_eq!(report.status(), reqwest::StatusCode::NO_CONTENT);
wait_for_session_activity(&h, id, ActivityState::Active).await;
let mut activities = Vec::new();
tokio::time::timeout(Duration::from_secs(2), async {
loop {
if let maestro_protocol::CatalogEvent::SessionUpdated { session } =
next_event(&mut ws).await
{
activities.push(session.activity);
if session.activity == ActivityState::Idle {
return;
}
}
}
})
.await
.expect("structured turn returns to idle");
assert!(
activities.contains(&ActivityState::Active),
"{activities:?}"
);
assert!(
!activities.contains(&ActivityState::Error),
"{activities:?}"
);
h.manager.kill(id, false).await.unwrap();
}
/// Each subscriber has its own lag domain now that the publisher task is gone:
/// two clients connected at different times both get their own Snapshot.
#[tokio::test]
+64 -16
View File
@@ -6,8 +6,8 @@ use std::collections::HashMap;
use std::path::PathBuf;
use chrono::{DateTime, Utc};
use maestro_daemon::sidecar::SidecarStore;
use maestro_protocol::{Session, SessionStatus, SessionView};
use maestro_daemon::sidecar::{PersistedEvent, SidecarStore};
use maestro_protocol::{Session, SessionStatus, SessionView, TranscriptRow, TranscriptRowKind};
use serde_json::Value;
use tempfile::TempDir;
use uuid::Uuid;
@@ -186,6 +186,13 @@ fn text_of(n: &SessionNotification) -> &str {
}
}
fn event_text(event: &maestro_daemon::sidecar::LoggedEvent) -> &str {
let maestro_daemon::sidecar::PersistedEvent::Notification(notification) = &event.event else {
panic!("expected ACP notification")
};
text_of(notification)
}
#[test]
fn append_assigns_line_position_as_seq_and_reads_back() {
let dir = TempDir::new().unwrap();
@@ -194,9 +201,9 @@ fn append_assigns_line_position_as_seq_and_reads_back() {
let mut log = store.open_events(id).expect("open");
assert_eq!(log.last_seq(), 0);
assert_eq!(log.append(&note("a")).unwrap(), 1);
assert_eq!(log.append(&note("b")).unwrap(), 2);
assert_eq!(log.append(&note("c")).unwrap(), 3);
assert_eq!(log.append(note("a")).unwrap(), 1);
assert_eq!(log.append(note("b")).unwrap(), 2);
assert_eq!(log.append(note("c")).unwrap(), 3);
assert_eq!(log.last_seq(), 3);
let path = store.events_path_for(id);
@@ -211,9 +218,7 @@ fn append_assigns_line_position_as_seq_and_reads_back() {
let all = store.read_events(id, EventCursor::Since(0), 100).unwrap();
assert_eq!(all.iter().map(|e| e.seq).collect::<Vec<_>>(), vec![1, 2, 3]);
assert_eq!(
all.iter()
.map(|e| text_of(&e.notification))
.collect::<Vec<_>>(),
all.iter().map(event_text).collect::<Vec<_>>(),
vec!["a", "b", "c"]
);
}
@@ -225,7 +230,7 @@ fn since_reads_forward_and_before_reads_backward_with_limit() {
let id = Uuid::new_v4();
let mut log = store.open_events(id).unwrap();
for i in 1..=6 {
log.append(&note(&format!("n{i}"))).unwrap();
log.append(note(&format!("n{i}"))).unwrap();
}
// `since` is exclusive: what a client that already holds seq 2 needs.
@@ -278,7 +283,7 @@ fn malformed_lines_are_skipped_but_keep_their_seq() {
let store = SidecarStore::new(dir.path());
let id = Uuid::new_v4();
let mut log = store.open_events(id).unwrap();
log.append(&note("first")).unwrap();
log.append(note("first")).unwrap();
drop(log);
// Garbage in the middle and a line that is JSON but not a notification.
@@ -289,11 +294,11 @@ fn malformed_lines_are_skipped_but_keep_their_seq() {
std::fs::write(&path, raw).unwrap();
let mut log = store.open_events(id).unwrap();
assert_eq!(log.last_seq(), 3);
assert_eq!(log.append(&note("fourth")).unwrap(), 4);
assert_eq!(log.append(note("fourth")).unwrap(), 4);
let all = store.read_events(id, EventCursor::Since(0), 10).unwrap();
assert_eq!(all.iter().map(|e| e.seq).collect::<Vec<_>>(), vec![1, 4]);
assert_eq!(text_of(&all[1].notification), "fourth");
assert_eq!(event_text(&all[1]), "fourth");
}
#[test]
@@ -303,7 +308,7 @@ fn truncated_last_line_still_opens_and_the_next_append_starts_a_fresh_line() {
let store = SidecarStore::new(dir.path());
let id = Uuid::new_v4();
let mut log = store.open_events(id).unwrap();
log.append(&note("whole")).unwrap();
log.append(note("whole")).unwrap();
drop(log);
let path = store.events_path_for(id);
@@ -319,10 +324,10 @@ fn truncated_last_line_still_opens_and_the_next_append_starts_a_fresh_line() {
// glue onto it; the partial keeps a seq of its own.
let mut log = store.open_events(id).unwrap();
assert_eq!(log.last_seq(), 2);
assert_eq!(log.append(&note("after")).unwrap(), 3);
assert_eq!(log.append(note("after")).unwrap(), 3);
let all = store.read_events(id, EventCursor::Since(0), 10).unwrap();
assert_eq!(all.iter().map(|e| e.seq).collect::<Vec<_>>(), vec![1, 3]);
assert_eq!(text_of(&all[1].notification), "after");
assert_eq!(event_text(&all[1]), "after");
let raw = std::fs::read_to_string(&path).unwrap();
assert_eq!(raw.lines().count(), 3, "{raw}");
}
@@ -334,8 +339,10 @@ fn delete_removes_the_event_log_and_load_all_ignores_it() {
let id = Uuid::new_v4();
store.write(&sample(id)).unwrap();
let mut log = store.open_events(id).unwrap();
log.append(&note("x")).unwrap();
log.append(note("x")).unwrap();
drop(log);
let legacy_approvals = dir.path().join(format!("{id}.approvals.jsonl"));
std::fs::write(&legacy_approvals, b"obsolete").unwrap();
assert!(store.events_path_for(id).exists());
let loaded = store.load_all().unwrap();
@@ -344,6 +351,47 @@ fn delete_removes_the_event_log_and_load_all_ignores_it() {
store.delete(id).unwrap();
assert!(!store.path_for(id).exists());
assert!(!store.events_path_for(id).exists());
assert!(!legacy_approvals.exists());
// Idempotent with the log already gone.
store.delete(id).unwrap();
}
#[test]
fn approval_rows_share_event_cursors_with_legacy_acp_lines() {
let dir = TempDir::new().unwrap();
let store = SidecarStore::new(dir.path());
let id = Uuid::new_v4();
let legacy = serde_json::to_vec(&note("before")).unwrap();
std::fs::create_dir_all(dir.path()).unwrap();
std::fs::write(store.events_path_for(id), [legacy, b"\n".to_vec()].concat()).unwrap();
let mut log = store.open_events(id).unwrap();
assert_eq!(
log.append(PersistedEvent::Approval {
row: TranscriptRow {
id: "approval:secret".into(),
kind: TranscriptRowKind::Approval,
data: serde_json::json!({"tool": "bash"}),
},
})
.unwrap(),
2
);
assert_eq!(log.append(note("after")).unwrap(), 3);
let all = store.read_events(id, EventCursor::Since(0), 10).unwrap();
assert_eq!(
all.iter().map(|event| event.seq).collect::<Vec<_>>(),
[1, 2, 3]
);
assert!(matches!(all[1].event, PersistedEvent::Approval { .. }));
assert_eq!(event_text(&all[0]), "before");
assert_eq!(event_text(&all[2]), "after");
let since = store.read_events(id, EventCursor::Since(1), 1).unwrap();
assert_eq!(since[0].seq, 2);
let before = store
.read_events(id, EventCursor::Before(Some(3)), 1)
.unwrap();
assert_eq!(before[0].seq, 2);
}
+51 -7
View File
@@ -156,9 +156,9 @@ pub enum SessionStatus {
Dead,
}
/// Inferred session activity, orthogonal to [`SessionStatus`] (NAP-340). The
/// daemon derives this from tmux pane state, PTY quiescence, and `/proc`
/// inspection — never by parsing the terminal byte stream.
/// Session activity, orthogonal to [`SessionStatus`] (NAP-340). Terminal
/// sessions use pane/process inference and agent reports; structured sessions
/// use their ACP turn lifecycle (docs/SIDEBAR.md).
///
/// Deserializes leniently (#110, SIDEBAR.md Precedence): an unknown variant
/// from a newer daemon becomes [`ActivityState::Idle`] instead of failing the
@@ -166,15 +166,14 @@ pub enum SessionStatus {
/// forgiven.
#[derive(Debug, Clone, Copy, PartialEq, Eq, Serialize, Default)]
pub enum ActivityState {
/// Producing output, or the foreground process is busy.
/// Producing output, the foreground process is busy, or an ACP turn is active.
Active,
/// Foreground process alive but quiescent. The forward-compatible default.
#[default]
Idle,
/// Blocked on a `read()` from the pty — waiting on the user.
/// Blocked on terminal input or waiting for an ACP approval.
AwaitingInput,
/// The agent running in the session reported a failure (#102). Never
/// inferred by the daemon; only set when an agent reports it (#105).
/// An agent reported a failure, or a structured session/turn failed.
Error,
}
@@ -794,6 +793,19 @@ pub struct TranscriptSnapshot {
pub rows: Vec<TranscriptRow>,
}
/// One durable, client-neutral replay step from a structured session.
///
/// Results are ordered by increasing `cursor`. `since` returns the first
/// matching steps; `before` returns the final matching steps, still increasing.
#[derive(Debug, Clone, PartialEq, Serialize, Deserialize)]
pub struct StructuredReplayEvent {
/// Durable ACP-log position represented by this projected step.
pub cursor: u64,
#[serde(skip_serializing_if = "Option::is_none")]
pub transcript: Option<TranscriptDelta>,
pub state: ReducedState,
}
/// A change to a transcript row.
#[derive(Debug, Clone, PartialEq, Serialize, Deserialize)]
#[serde(tag = "type", rename_all = "snake_case")]
@@ -836,6 +848,38 @@ pub struct ReducedState {
pub mode: Option<String>,
}
/// Server messages sent over a structured-session WebSocket.
#[derive(Debug, Clone, PartialEq, Serialize, Deserialize)]
#[serde(tag = "type", rename_all = "snake_case")]
pub enum StructuredEvent {
TranscriptSnapshot {
snapshot: TranscriptSnapshot,
},
ReducedState {
state: ReducedState,
},
ProjectionUpdate {
#[serde(skip_serializing_if = "Option::is_none")]
transcript: Option<TranscriptDelta>,
state: ReducedState,
},
}
/// HTTP body for `POST /api/sessions/:id/structured/prompt`.
#[derive(Debug, Clone, PartialEq, Eq, Serialize, Deserialize)]
pub struct StructuredPromptRequest {
pub prompt: String,
}
/// Operator decision for a pending ACP permission request.
#[derive(Debug, Clone, Copy, PartialEq, Eq, Serialize, Deserialize)]
#[serde(rename_all = "snake_case")]
pub enum StructuredApprovalDecision {
AllowOnce,
AllowForSession,
Deny,
}
// ---------------------------------------------------------------------------
// Client ↔ Daemon control messages (text frames on the attach WS)
// ---------------------------------------------------------------------------
+66
View File
@@ -71,6 +71,31 @@ pub fn attach(id: Uuid) -> String {
format!("{API_PREFIX}/attach/{id}")
}
/// `WS /api/structured/:id` — transcript/state snapshots and live deltas.
pub fn structured(id: Uuid) -> String {
format!("{API_PREFIX}/structured/{id}")
}
/// `GET /api/sessions/:id/structured/replay` — cursor-bounded ACP history.
pub fn structured_replay(id: Uuid) -> String {
format!("{SESSIONS}/{id}/structured/replay")
}
/// `POST /api/sessions/:id/structured/prompt` — send an ACP prompt.
pub fn structured_prompt(id: Uuid) -> String {
format!("{SESSIONS}/{id}/structured/prompt")
}
/// `POST /api/sessions/:id/structured/cancel` — cancel an ACP turn.
pub fn structured_cancel(id: Uuid) -> String {
format!("{SESSIONS}/{id}/structured/cancel")
}
/// `POST /api/sessions/:id/structured/approvals/:nonce` — answer ACP permission.
pub fn structured_approval(id: Uuid, nonce: &str) -> String {
format!("{SESSIONS}/{id}/structured/approvals/{nonce}")
}
/// Axum-style patterns for the two parameterised routes, so the server
/// registers the same paths [`session`] and [`attach`] build.
pub mod patterns {
@@ -80,6 +105,11 @@ pub mod patterns {
pub const SESSION_PROMPT: &str = "/api/sessions/:id/prompt";
pub const SESSION_FOCUS: &str = "/api/sessions/:id/focus";
pub const ATTACH: &str = "/api/attach/:id";
pub const STRUCTURED: &str = "/api/structured/:id";
pub const STRUCTURED_REPLAY: &str = "/api/sessions/:id/structured/replay";
pub const STRUCTURED_PROMPT: &str = "/api/sessions/:id/structured/prompt";
pub const STRUCTURED_CANCEL: &str = "/api/sessions/:id/structured/cancel";
pub const STRUCTURED_APPROVAL: &str = "/api/sessions/:id/structured/approvals/:nonce";
}
/// A daemon base URL that could not be turned into a WebSocket base.
@@ -206,6 +236,22 @@ mod tests {
attach(id()),
"/api/attach/6f1a7c0e-0f5e-4a2b-9a3d-6b1c2d3e4f50"
);
assert_eq!(
structured(id()),
"/api/structured/6f1a7c0e-0f5e-4a2b-9a3d-6b1c2d3e4f50"
);
assert_eq!(
structured_replay(id()),
"/api/sessions/6f1a7c0e-0f5e-4a2b-9a3d-6b1c2d3e4f50/structured/replay"
);
assert_eq!(
structured_prompt(id()),
"/api/sessions/6f1a7c0e-0f5e-4a2b-9a3d-6b1c2d3e4f50/structured/prompt"
);
assert_eq!(
structured_cancel(id()),
"/api/sessions/6f1a7c0e-0f5e-4a2b-9a3d-6b1c2d3e4f50/structured/cancel"
);
assert_eq!(
attach_url("https://laptop.lan:7070", id()).unwrap(),
"wss://laptop.lan:7070/api/attach/6f1a7c0e-0f5e-4a2b-9a3d-6b1c2d3e4f50"
@@ -240,6 +286,22 @@ mod tests {
patterns::ATTACH.trim_end_matches(":id"),
attach(id()).trim_end_matches(&id().to_string())
);
assert_eq!(
patterns::STRUCTURED.trim_end_matches(":id"),
structured(id()).trim_end_matches(&id().to_string())
);
assert_eq!(
patterns::STRUCTURED_REPLAY.trim_end_matches(":id/structured/replay"),
session(id()).trim_end_matches(&id().to_string())
);
assert_eq!(
patterns::STRUCTURED_PROMPT.trim_end_matches(":id/structured/prompt"),
session(id()).trim_end_matches(&id().to_string())
);
assert_eq!(
patterns::STRUCTURED_CANCEL.trim_end_matches(":id/structured/cancel"),
session(id()).trim_end_matches(&id().to_string())
);
}
#[test]
@@ -258,5 +320,9 @@ mod tests {
assert!(path.starts_with(API_PREFIX), "{path} escapes {API_PREFIX}");
}
assert!(attach(id()).starts_with(API_PREFIX));
assert!(structured(id()).starts_with(API_PREFIX));
assert!(structured_replay(id()).starts_with(API_PREFIX));
assert!(structured_prompt(id()).starts_with(API_PREFIX));
assert!(structured_cancel(id()).starts_with(API_PREFIX));
}
}
+15 -5
View File
@@ -18,10 +18,14 @@ wizard, and cleans up — only the unified `maestro` binary ends up persisted in
**Updating is the same command.** Re-running `./install.sh` detects an existing
systemd unit, refreshes it, and runs `systemctl --user restart maestro` instead
of `start`. No `--update` flag needed. **The restart keeps your sessions:** the unit sets
of `start`. No `--update` flag needed. **The restart keeps your terminal sessions:** the unit sets
`KillMode=process`, so systemd signals only the daemon — the tmux server (and
every session) survives, and the new daemon re-adopts them on start. A unit
written before this default lacks the line; re-run `./install.sh` to refresh it.
every tmux-backed session) survives, and the new daemon re-adopts them on
start. Structured sessions (an agent run as a daemon child over ACP) are not
tmux panes, so shutdown ends their active turn; the transcript survives and
the next daemon attempts one resume (`docs/RESTART.md#structured-sessions`). A
unit written before this default lacks the line; re-run `./install.sh` to
refresh it.
Inside the script, `cargo install --git https://git.naps.pt/naps62/maestro …`
uses your already-configured git credentials for this host (HTTPS helper or
@@ -94,7 +98,10 @@ running a local daemon, for a host that only talks to another machine's.
The daemon unit carries the same `KillMode=process` as the wizard-written one,
so `systemctl --user restart maestro-daemon` leaves the tmux server and every
session alive.
terminal session alive. Structured sessions end their active turn on
shutdown and resume once on the next daemon: ACP state comes back via
`session/load`, while the event log only replays Maestro's own transcript
projection (`docs/RESTART.md#structured-sessions`).
## Thin builds for single-role hosts
@@ -211,7 +218,10 @@ what to run. The adapter is a child of the daemon: it ends with a `DELETE
itself. A daemon restart ends an in-flight structured turn, but preserves its
worktree, session record, and event-log transcript. The replacement daemon
starts a new adapter, resumes it with `session/load`, and replays the
transcript. Terminal sessions keep surviving restarts through tmux.
transcript. Terminal sessions keep surviving restarts through tmux. See
`docs/STRUCTURED-VIEW.md` for the full contract — row kinds, projections, the
handshake, and the approval and recovery rules — and `docs/RESTART.md` for
what a client observes across the restart.
## The browser UI
+39 -4
View File
@@ -2,10 +2,16 @@
Maestro is upgraded in place: the operator reinstalls the binary and systemd
restarts the daemon under every browser tab and TUI that is attached to it.
tmux keeps the sessions alive across that (see `docs/INSTALL.md`), but clients
lose their sockets. This document is the contract for what a client does about
it — what the daemon publishes, how a client tells a restart from a blip, and
what the user sees while it happens.
tmux keeps **terminal** sessions alive across that (see `docs/INSTALL.md`),
but clients lose their sockets either way. This document is the contract for
what a client does about it — what the daemon publishes, how a client tells a
restart from a blip, and what the user sees while it happens.
A **structured** session (`docs/STRUCTURED-VIEW.md`) has no tmux pane behind
it to survive on: its agent is a child of the daemon process itself, so a
restart takes it down with the daemon. See
[Structured sessions](#structured-sessions) below for exactly what that
costs and what it doesn't.
## What the daemon publishes
@@ -68,6 +74,35 @@ resolves, so a blocked client is never a trapped one. It clears itself the
moment a probe succeeds — there is no "retry" button, because the client is
already retrying.
## Structured sessions
Everything above describes a **client's** view of a restart, and it holds for
a structured session exactly as written — the same three verdicts, the same
status indicator, the same overlay. What differs is what survives on the
**daemon** side, because a structured session's agent is a daemon child, not
a tmux pane:
- A restart **ends the active turn**. There is no adapter process left to
finish it, and no tmux pane keeping it alive underneath.
- The **transcript survives**, in the per-session event log
(`docs/STRUCTURED-VIEW.md#persistence-recovery-and-dead-sessions`) — a
graceful shutdown marks the session for one resume attempt. Recovery is
two separate operations, not one: the next daemon calls `session/load` so
the adapter restores its own ACP session, but whatever history the
adapter emits during that call is discarded, not replayed into anything —
the daemon already has that history on disk. Separately, the daemon
rebuilds its own projection by folding the persisted event log back
through it. An ungraceful death (a crash, `kill -9`) leaves no resume
marker, and the session stays dead permanently.
- **Pending approvals and their nonces do not survive.** Both live only in
the old daemon's memory; any ACP permission request still mid-flight when
the daemon exits is answered `Cancelled` on the way down, and a resumed
session starts with no pending approval at all.
`docs/STRUCTURED-VIEW.md` is the full contract for all of this — the event
log's shape, the recovery marker, and exactly which kind of session death is
resumable.
## Timing
- **Grace window** — 1200 ms between the socket dropping and the overlay
+22 -5
View File
@@ -231,7 +231,24 @@ cutover points land within one tick.
## Precedence
An agent-reported `ActivityState` (#105) overrides the daemon's inference for
### Activity source by session kind
The daemon supplies one `activity` field to both clients, but its source depends
on the session kind:
| Session kind | Source |
|---|---|
| Terminal | The existing agent-report override and five-minute TTL, falling back to PTY quiescence, foreground process and `/proc` input-block inference. |
| Structured | ACP state directly: a pending approval is `AwaitingInput`; otherwise an in-flight turn is `Active`; no turn is `Idle`. A failed turn or dead structured session is `Error`, and the next turn clears a failed-turn error. |
Structured sessions never enter the terminal inference path and ignore activity
reports inherited through agent hooks. They have no PTY whose output or blocked
read could describe their state. This is a second source, not a replacement for
the terminal source.
### Terminal report override
For terminal sessions, an agent-reported `ActivityState` (#105) overrides the daemon's inference for
**five minutes** (the TTL is fixed by the daemon; re-reporting extends it).
When the TTL expires, inference resumes and the session's reported detail —
if any — clears with the report. Clients never blend the two: whichever source
@@ -281,10 +298,10 @@ session is genuinely `Active` again — new work, not merely a refocus. Only
after that fresh `Active` cycle does the *next* `AwaitingInput` glow again.
Unfocusing and refocusing alone never re-arms it.
This applies uniformly to `AwaitingInput` from either source — the daemon's
own inference or an agent's report (#105) — since the rail draws them
identically. It does not touch `Error`; that state's own TTL (below) is
unrelated and unaffected by focus.
This applies uniformly to `AwaitingInput` from terminal inference, an agent
report (#105), or a structured session's pending approval, since the rail draws
them identically. It does not touch `Error`; terminal reports keep their TTL,
while a structured failed-turn error clears when the next turn starts.
Each client reports its own pane's focus with `POST /api/sessions/:id/focus`
(`{"focused": bool}`) whenever the pane it is showing changes — on attach, on
+387
View File
@@ -0,0 +1,387 @@
# STRUCTURED-VIEW.md — the structured session contract
The contract for a structured session (`view: Structured`, #133): an agent run
as a daemon child over the Agent Client Protocol (ACP) instead of in a tmux
pane. Alongside `docs/WORKTREES.md`, `docs/SIDEBAR.md` and `docs/RESTART.md`,
read this before changing anything that renders a structured session — both
the web and TUI renderers are written against it.
It covers the eight transcript row kinds, what the daemon decides versus what
a client decides, the two projections and their delta/cursor semantics, the
ACP handshake and the capabilities it declines, the adapter's install
command, and the approval nonce rules. **Non-goal: the detached runner.** It
does not exist yet, and nothing here describes it.
## What a structured session is
A terminal session's agent runs inside a tmux pane; a structured session's
agent runs as a subprocess of the daemon itself, speaking ACP over stdio. The
daemon is the ACP **client**; the adapter process, `claude-code-acp`, is the
ACP **agent** — it wraps Claude Code's own SDK and translates the model's
tool calls into ACP `session/update` notifications instead of drawing them to
a pane. `crates/maestro-acp` owns the daemon's side of that connection;
`crates/maestro-daemon/src/projection.rs` folds what it receives into the two
projections below.
The child is a real child of the daemon process: same process group, and on
Linux marked with `PR_SET_PDEATHSIG` so it dies with the daemon even if the
daemon exits without cleaning up; other platforms have no such signal and
depend on an orderly shutdown/drop to terminate it. The ordinary path on
every platform is `AcpSession::terminate` (`session/cancel`, a short grace,
`SIGTERM`, then `SIGKILL`), called from both `DELETE /api/sessions/:id` and
daemon shutdown.
`docs/WORKTREES.md`'s kill-guard rule — an unforced kill refused while the
session's worktree holds unsaved work — applies to a structured session
exactly as it does to a terminal one; only what gets terminated differs (an
ACP handle instead of a tmux pane).
## The adapter
`ADAPTER_BIN` (`claude-code-acp`) is resolved by walking `PATH` — the
session's own `PATH` env var if the spawn request set one, the daemon's
otherwise — and taking the first executable file named `claude-code-acp` on
it. Install it once per host:
```sh
npm install -g --prefix ~/.local @zed-industries/claude-code-acp
```
That is `maestro_acp::INSTALL_COMMAND`, and it is not just documentation: it
is baked into `AcpError::AdapterMissing`'s message. When the adapter cannot
be found, spawn still creates the session — **Dead**, with that install
command as its `activity_detail` — so the sidebar names the fix instead of
the session silently never appearing (see `docs/INSTALL.md` for the full
install writeup, including the systemd `PATH` caveat). The adapter also needs
a logged-in Claude Code (`claude /login`); a missing login surfaces the same
way, via the adapter's own stderr tail in the failure message.
The handshake is bounded by `HANDSHAKE_TIMEOUT` (60s) — long enough for a
slow machine to boot a node process and the Claude Code SDK behind it, never
long enough to hang a spawn indefinitely. Any handshake failure — missing
adapter, early exit, a protocol error, or the timeout — leaves no orphaned
process: `spawn` closes and reaps the child itself before returning the
error.
## The ACP handshake
`initialize` is where the daemon states what kind of client it is. It
declares, deliberately, almost nothing:
| Capability | Declared | Why |
| --- | --- | --- |
| `fs.readTextFile` | `false` | The adapter uses Claude Code's own Read tool instead of asking the daemon to read a file on its behalf — the same tool a terminal-view Claude Code session uses. |
| `fs.writeTextFile` | `false` | Same reasoning, for Write. |
| `terminal` | `false` | The adapter runs Claude Code's own Bash tool rather than asking the daemon to open a terminal for it. |
Declining all three keeps the daemon out of the file/terminal surface
entirely — the adapter already has direct access to the session's `cwd`, and
routing every read, write, and shell command through the daemon would only
add a hop with no new capability. Every other capability namespace ACP
defines (session extensions, auth, elicitation) is simply left at its
default; the daemon does not advertise support because `claude-code-acp` has
never asked for it, not because it was weighed and refused.
The one client method that *is* implemented for real is
`session/request_permission` (#139) — see [Approvals](#approvals) below.
Before #139 it was a stub that declined every request; that stub is gone now
that a live implementation exists.
`session/new` opens a fresh ACP session on every ordinary spawn. A respawn
after a daemon restart instead sets `SpawnConfig::load_session(id)` and runs
`session/load` with the ACP session id the original handshake returned
(`Session::acp_session_id` on the wire) — the adapter must advertise
`agent_capabilities.load_session`, or the handshake fails outright naming
that. Whatever history the adapter streams back while `session/load` is in
flight is dropped on arrival, deliberately: the daemon's own event log
already holds that history, and forwarding it too would duplicate every row
in the projections below.
## The eight transcript row kinds
`TranscriptRowKind` (`maestro-protocol`) is the alphabet a structured
transcript is written in. Every row has a stable `id` (so later deltas can
address it), a `kind`, and a `data: Value` that preserves the underlying ACP
payload — clients depend on ACP's *shape* without depending on ACP as a
crate.
| Kind | Carries | Row identity |
| --- | --- | --- |
| `User` | An echoed `user_message_chunk` — the prompt text as ACP saw it. | `user:<message_id>`, or `user:<seq>` when the chunk carries no id. |
| `AgentText` | An `agent_message_chunk` — the agent's visible reply, streamed. | `agent:<message_id>` / `agent:<seq>`. Later chunks for the same id are appended into `data.content.text` in place, not appended as new rows. |
| `AgentThought` | An `agent_thought_chunk` — the model's reasoning stream. | `thought:<message_id>` / `thought:<seq>`, coalesced the same way as `AgentText`. |
| `ToolCall` | A `tool_call` / `tool_call_update` pair whose content carries no diff. | `tool:<tool_call_id>` — one row for the call's whole life; updates patch it in place, they never create a second row. |
| `ToolDiff` | The same tool-call machinery, when the tool's content includes a `Diff` block. | Same `tool:<tool_call_id>` identity as `ToolCall`; the *kind* flips between the two as the tool's own content changes, not the row. |
| `Approval` | A daemon-originated permission request (`RequestPermissionRequest`), serialized without its nonce. | `approval:<uuid>` — minted fresh per request, unrelated to the ACP request's own id. |
| `Plan` | A `plan` update — the agent's current step list. | The fixed id `"plan"`. Each new plan **replaces** the row wholesale; entries are not diffed against the previous plan. |
| `Notice` | The raw JSON of any `session/update` variant the ACP dependency already decodes but the projection does not special-case. | `notice:<seq>`. |
`Notice` covers known-but-unspecialized `SessionUpdate` variants: ones the
`agent-client-protocol` dependency's schema already deserializes (it
currently folds `available_commands_update` and `session_info_update`
there) but the projection's `match` doesn't fold into one of the other
seven kinds. It is not a wire-level forward-compatibility seam:
`SessionUpdate` is an internally-tagged enum with no unknown-variant
fallback, so a `session/update` whose `sessionUpdate` tag this build's
dependency has never heard of — including a variant gated behind an
unstable ACP feature this build doesn't enable, which is compiled out of
the enum entirely — fails to deserialize before the projection ever runs,
rather than becoming a `Notice`. A genuinely new ACP variant needs the
`agent-client-protocol` dependency upgraded (and, typically, the projection's
`match` taught about it) before it can appear on this transcript at all — see
`projection.rs`'s test asserting that an unrecognised `sessionUpdate` tag is a
deserialization error, not a `Notice`. A client that does not know what to do
with a `Notice` can still render it as an opaque, collapsed line — that
remains a reasonable floor, not a client bug.
Rendering is entirely a client concern. The projection's job stops at
folding ACP into these eight shapes; markdown formatting, diff rendering,
tool-card treatment, and how a `Notice` is presented are all renderer
choices, not daemon ones.
## Daemon decisions vs. client decisions
**The daemon decides:**
- Which ACP `session/update` becomes which row kind, and under what id — the
fold in `projection.rs` is the one place this logic lives, so the web and
TUI renderers can never disagree about it.
- Turn lifecycle: when a turn begins, and whether an ACP response, a cancel,
or a failure ends it (`TurnState.responding` / `.thinking` /
`.active_tool_calls` are derived, not renderer-computed).
- Approval nonce minting, secrecy, and the single-active-approval rule (see
below) — a client never sees an ACP permission request before the daemon
has already turned it into an `Approval` row and a nonce.
- What happens to an ACP update variant the dependency decodes but nobody has
taught the projection to special-case (falls back to `Notice` rather than
erroring). A variant the dependency itself can't decode is a deserialization
error, not a `Notice` — see [The eight transcript row
kinds](#the-eight-transcript-row-kinds).
- Persistence and replay: what is written to the event log, in what order,
and how a restart resumes it.
**A client decides:**
- How each row kind is drawn — markdown for `AgentText`/`AgentThought`, a
diff view for `ToolDiff`, a tool card for `ToolCall`, a checklist for
`Plan`, and a floor presentation for `Notice`.
- How `responding` vs. `thinking` are distinguished visually, if at all.
- How the operator is asked to resolve an `Approval` row — the daemon only
exposes the three-way decision (`AllowOnce` / `AllowForSession` / `Deny`),
never the UI that solicits it.
- Scrolling, collapsing, and every other presentation-only behaviour.
## The two projections
`Projection` (daemon-only) is folded once from the durable ACP stream and
exposes two independent views, both defined in `maestro-protocol`:
- **`TranscriptSnapshot { cursor, rows }`** — the complete ordered list of
transcript rows built so far, plus the log position (`cursor`) it reflects.
It is intentionally always the *whole* transcript, not a windowed one: a
tool row patched much later must still find its original entry.
- **`ReducedState`** — one whole-session snapshot folded over every event
seen: `turn` (the per-turn indicators above), `pending_approval`, `plan`,
`usage`/`context`, `model`, and `mode`. Like `TranscriptSnapshot`, it is
always full-history, never a delta — a client that only wants "is a turn
active right now" never has to replay anything to answer that.
Changes to the transcript travel as `TranscriptDelta`, tagged `type`:
`Append { row }` (a new row), `Patch { row }` (an existing row, addressed by
`id`, replaced wholesale), and `Remove { id }`. `Remove` is defined on the
wire for forward compatibility; nothing in the current projection emits it —
every row, once appended, is only ever patched or left alone.
### Cursor and replay semantics
The event log's own line position is the cursor: 1-based, shared by ACP
notifications and daemon-generated rows alike (`sidecar.rs`'s
`PersistedEvent` envelope is what lets an `Approval` row share a cursor
sequence with raw ACP JSON without pretending to be ACP traffic). Two ways to
read it:
- **`WS /api/structured/:id`** — on connect, sends a `TranscriptSnapshot` and
a `ReducedState` covering the whole log as it stands, then streams
`ProjectionUpdate { transcript, state }` for everything that happens next.
There is no client-supplied resume cursor on this socket: every connection
gets the full snapshot, not a gap-filled continuation, so a client never
needs the replay endpoint just to obtain the current transcript. Use
replay instead for cursor-bounded pagination through the durable log —
history from before this connection, fetched a page at a time.
- **`GET /api/sessions/:id/structured/replay`** — cursor-bounded pagination
over the durable log, returning one `StructuredReplayEvent` (`cursor`,
`transcript`, `state`) per matching line. `?since=<n>` returns the first
page after `n` (`since=0` selects from the beginning, i.e. the first
page); a client pages forward from there using the cursor of the last
event returned. `?before=<n>` returns the page immediately before `n`,
still in ascending order — passing both is a `400`. `?limit` defaults to
100 and is clamped to 500. Each call re-folds
the entire log from scratch to compute the state at every returned cursor,
so a page reflects exactly the *durable* state at its cursor, not an
approximation of it — at the cost of redoing the whole fold per request.
That durable state is not always identical to what a live client saw at
that moment, though: turn lifecycle transitions are broadcast live without
a durable event backing them (see [Prompt and cancel](#prompt-and-cancel-bounded-liveness)),
and a restored `Approval` row carries no pending nonce, because the nonce
never reached disk (see [Approvals](#approvals)) — replay is exact and
reconstructible, not a perfect record of everything the live view showed.
A session that is not a live structured session (terminal, or a dead
adapter) answers `409 not_structured` from either route rather than a
coherent-looking but stale reply.
## Approvals
An ACP `session/request_permission` callback becomes an `Approval` row plus a
pending decision, entirely on the daemon side:
- **Generation.** The daemon mints the nonce itself — a fresh
`Uuid::new_v4()` string per request — rather than using anything ACP's own
request carries. ACP permission callbacks have no request id of their own
that would double as one.
- **Live-only placement.** The nonce lives in exactly one place: the
in-memory `pending_approvals` map keyed by nonce, and its mirror in
`ReducedState.pending_approval.nonce` on the live wire. It is never part of
the `Approval` row's `data` — the row is `RequestPermissionRequest`
serialized *without* the nonce.
- **Secrecy from the agent and from durable history.** The adapter itself
never learns the nonce — it is a daemon-to-client secret, never sent back
over the ACP connection. It is likewise never written to the event log:
restoring an `Approval` row from disk (`Projection::restore_approval`)
produces a historical row with no `pending_approval` in state, because the
nonce that would make it "pending" never made it to disk in the first
place. A restart therefore cannot resume a pending approval — see
[Persistence](#persistence-recovery-and-dead-sessions).
- **One live approval at a time.** A new permission request mints its own
nonce, then cancels whatever was already pending (`cancel_pending_approvals`)
before installing itself as the new pending approval — ACP serializes
requests through one connection, so this only ever fires if an adapter
nonetheless overlaps two. Either way, the prior pending request is
cancelled/cleared before the new one is installed, so the newest request
wins.
- **Single-use and duplicate resolution.**
`POST /api/sessions/:id/structured/approvals/:nonce` **always answers
`204`** for a known, live structured session — whether the nonce is the
one currently pending, one already consumed by an earlier answer, or one
that never existed. Only an unknown or non-structured *session* answers
otherwise (`404` / `409`). This is deliberate: two clients racing to answer the same
approval must not turn a harmless duplicate click into a client-visible
error. Whichever `POST` reaches the daemon first removes the nonce from
`pending_approvals` and is the one whose decision actually reaches the
adapter; every later `POST` for that nonce is a no-op `204`.
- **Cancellation vs. denial.** A turn cancel or a session shutdown resolves
any pending approval with ACP's `Cancelled` outcome, not a `Deny` decision
— the adapter should read those as "the operation was withdrawn," not "the
operator refused it."
- **Decision mapping.** `AllowOnce` and `AllowForSession` select the ACP
option of kind `AllowOnce` / `AllowAlways` from the request's own option
list. `Deny` prefers a one-shot refusal (`RejectOnce`) and falls back to a
permanent one (`RejectAlways`) only when that is the sole rejection option
offered; if neither is present, it answers `Cancelled`.
## Prompt and cancel: bounded liveness
`POST /api/sessions/:id/structured/prompt` and `.../cancel` are serialized
against each other through one per-session `operation` lock
(`AcpEventStream::operation`), because ACP permission callbacks carry no
per-request id — the daemon's own notion of "the active turn" is the only
thing a callback can be attributed to, so at most one turn may ever be
in flight, and a cancelled turn must fully settle before the next prompt is
allowed to claim the daemon's idea of "the active turn."
- A prompt claims the turn and dispatches it to ACP as one operation. While a
turn is active, a second prompt gets `409 turn_in_flight` rather than
queuing — one active turn per session, never a hidden queue of prompts.
- Cancel sends `session/cancel`, clears any pending approval as `Cancelled`,
and ends the turn's client-visible lifecycle immediately. If the cancelled
turn was still *dispatched* to ACP (waiting on its response), the daemon
cannot simply forget it: ACP has not confirmed the cancel yet, and a new
prompt claiming "the active turn" while the old one is still outstanding
would make the next permission callback ambiguous between the two.
- So a cancelled-but-still-dispatched turn must **settle within
`CANCEL_GRACE` (500ms)** — either ACP acknowledges the prompt (however it
resolves) or that grace period runs out. If it settles in time, the next
prompt proceeds normally. If it does not, the daemon does not guess: it
marks the whole structured session dead (`fail_cancelled_acp_turn`) rather
than risk a callback resolving against the wrong turn. This is the bounded
half of "bounded liveness" — a structured session's prompt/cancel dance
either resolves promptly or the session is declared unusable, never left
in an indefinite ambiguous state.
## Persistence, recovery, and dead sessions
A structured session's durable state is three files in the sidecar
directory: the sidecar itself (`<id>.json`), plus two next to it:
- **`<id>.events.jsonl`** — the unified event log. Each line is either a raw
ACP `SessionNotification` (kept in its original JSON so old logs stay
readable) or a `{"maestroEvent": "approval", "row": …}` envelope for a
daemon-generated row. Both share the same 1-based line-position cursor. A
line a hard kill left half-written is terminated before the next append, so
it keeps its position and is skipped as malformed rather than corrupting
what follows. An append that fails *while the daemon is running* — a
`write_all` that errors after only a prefix reached disk, or a `flush` that
errors after a complete write — is rolled back to the exact byte length the
file was at before the call, so `last_seq` and the file's line count never
disagree and the same log handle can be appended to again immediately (#184).
If that rollback itself cannot restore the prior length, the append is
fatal: the log's consistency can no longer be promised, so the caller kills
the session instead of appending to it again.
- **`<id>.recover`** — an empty marker file. Its only meaning is "the daemon
that wrote this shut down gracefully while this session was running, and
the next daemon may attempt to resume it." It is consumed (deleted) the
moment a recovery attempt reads it — one recovery attempt per graceful
shutdown, never a standing flag.
Three ways a structured session ends up `Dead`, each with different
persistence consequences:
1. **The adapter exits, or a notification append to the event log fails,
while the daemon keeps running.** (An approval-row append failure is
narrower: as long as the log rolled back cleanly, it only cancels the
pending approval and rolls the unpersisted `Approval` row back out of the
live transcript, without marking the session dead — a snapshot taken after
the failure must never show a row durable replay could not reproduce
(#184). If the rollback itself did not restore the log's prior length,
this narrower path does not apply either: the session is marked dead the
same as any other append failure, since the log can no longer promise
`last_seq` matches its file.)
`mark_session_dead` applies the same natural-exit worktree rule
as a terminal session's own pane exiting (`docs/WORKTREES.md`): a pristine
worktree is removed, one holding unsaved work is left on disk. No recovery
marker is written. This death is permanent — a later daemon restart's
`recover()` finds the sidecar, sees no marker, and leaves the session
`Dead` without attempting to resume it.
2. **A cancelled turn fails to settle within `CANCEL_GRACE`.** Same
consequence as (1): the session is marked dead through the same path, no
marker, no future resumption.
3. **Graceful daemon shutdown**, while the session was still `Running`.
`mark_session_dead_after_shutdown` marks the session dead **without**
applying the worktree exit rule — the tree is kept exactly as it is,
pristine or not, because there is nothing "natural" about this exit — and
writes the `.recover` marker. The sidecar on disk therefore says `Dead`,
but is eligible for resumption.
Recovery, on the next daemon's `recover()`: for every `Structured` sidecar
not already adopted, the marker is consumed. When one was present and the
sidecar carries a stored `acp_session_id`, the daemon spawns a fresh adapter
with `SpawnConfig::load_session(acp_session_id)`, replays the
entire durable event log into a new `Projection` (rebuilding the transcript
and reduced state exactly as they were, with any restored `Approval` row
landing as history — never as a re-armed pending approval, since its nonce
never reached disk), and only then reopens the log for further appends. What
the adapter streams back while its own `session/load` is in flight is
discarded rather than appended, so the replayed history is not duplicated
(see [The ACP handshake](#the-acp-handshake)). Success flips the session back
to `Running`/`Idle`; failure (no stored `acp_session_id`, or the handshake
itself failing) leaves it `Dead` with the failure reason as
`activity_detail` — the same "tell the operator what to fix" pattern as a
missing adapter binary.
Pending approvals and their nonces never survive any of this. They live only
in the in-memory `pending_approvals` map and the live `ReducedState`, both
gone the instant the daemon that held them exits. Whatever ACP permission
request was mid-flight when the old daemon died was already answered
`Cancelled` on that daemon's own shutdown path, so a resumed session always
starts with no pending approval.
See `docs/RESTART.md` for what a *client* observes across a daemon restart,
now that structured sessions exist alongside terminal ones.