Files
codex/codex-rs/analytics/src/client_tests.rs
T
Shijie Rao 7abfcf220b Expose thread-level multi-agent mode (#28792)
## Why

Once multi-agent mode can be selected per turn, clients also need to
choose the initial selection when creating a thread and observe that
selection through lifecycle and settings APIs.

The selected value is intentionally distinct from the effective
model-visible value: no client selection is represented as `null`, even
though an eligible multi-agent v2 turn derives `explicitRequestOnly` as
its effective default.

## What changed

- Add the optional experimental `thread/start.multiAgentMode` parameter
and pass it through thread creation.
- Preserve an omitted initial value as an unset selection rather than
eagerly storing `explicitRequestOnly`.
- Apply an explicit `thread/start` selection to the first turn through
the session configuration established at thread creation.
- Restore the latest persisted effective mode as the selected baseline
on cold resume when rollout history contains one.
- Inherit the optional selected mode from a loaded parent when creating
related runtime threads.
- Return the current selected `multiAgentMode` from `thread/start`,
`thread/resume`, `thread/fork`, and thread settings, using `null` when
no mode is selected.
- Keep lifecycle reporting independent from model capability and feature
eligibility; core turn construction remains responsible for calculating
and persisting the effective mode.

## Not covered

- Clearing an existing loaded-session selection back to unset through
`turn/start`; omitted or `null` currently retains the session's
selection.
- A TUI control, slash command, or `config.toml` preference.

## Verification

- `CARGO_INCREMENTAL=0 just test -p codex-app-server-protocol`
- `CARGO_INCREMENTAL=0 just test -p codex-app-server multi_agent_mode`

The focused app-server coverage verifies explicit `thread/start`
initialization, first-turn prompting, nullable reporting for an omitted
selection, and retention of selections that are not currently
runtime-eligible.

## Stack

Stacked on #28685. This PR contains only the thread initialization and
lifecycle/settings API layer.
2026-06-19 10:50:44 +02:00

448 lines
16 KiB
Rust

use super::AnalyticsEventsClient;
use super::AnalyticsEventsDestination;
use super::AnalyticsEventsQueue;
#[cfg(debug_assertions)]
use super::capture_track_events_request;
#[cfg(debug_assertions)]
use super::send_track_events_request;
use super::track_event_request_batches;
use crate::events::CodexAcceptedLineFingerprintsEventParams;
use crate::events::CodexAcceptedLineFingerprintsEventRequest;
use crate::events::SkillInvocationEventParams;
use crate::events::SkillInvocationEventRequest;
use crate::events::TrackEventRequest;
use crate::facts::AnalyticsFact;
use crate::facts::InvocationType;
use codex_app_server_protocol::ApprovalsReviewer as AppServerApprovalsReviewer;
use codex_app_server_protocol::AskForApproval as AppServerAskForApproval;
use codex_app_server_protocol::ClientRequest;
use codex_app_server_protocol::ClientResponsePayload;
use codex_app_server_protocol::RequestId;
use codex_app_server_protocol::SandboxPolicy as AppServerSandboxPolicy;
use codex_app_server_protocol::SessionSource as AppServerSessionSource;
use codex_app_server_protocol::Thread;
use codex_app_server_protocol::ThreadArchiveParams;
use codex_app_server_protocol::ThreadArchiveResponse;
use codex_app_server_protocol::ThreadForkResponse;
use codex_app_server_protocol::ThreadResumeResponse;
use codex_app_server_protocol::ThreadStartResponse;
use codex_app_server_protocol::ThreadStatus as AppServerThreadStatus;
use codex_app_server_protocol::Turn;
use codex_app_server_protocol::TurnStartParams;
use codex_app_server_protocol::TurnStartResponse;
use codex_app_server_protocol::TurnStatus as AppServerTurnStatus;
use codex_app_server_protocol::TurnSteerParams;
use codex_app_server_protocol::TurnSteerResponse;
use codex_utils_absolute_path::test_support::PathBufExt;
use codex_utils_absolute_path::test_support::test_path_buf;
use std::collections::HashSet;
#[cfg(debug_assertions)]
use std::fs;
#[cfg(debug_assertions)]
use std::path::PathBuf;
use std::sync::Arc;
use std::sync::Mutex;
#[cfg(debug_assertions)]
use std::time::SystemTime;
use tokio::sync::mpsc;
use tokio::sync::mpsc::error::TryRecvError;
fn sample_accepted_line_fingerprint_event(thread_id: &str) -> TrackEventRequest {
TrackEventRequest::AcceptedLineFingerprints(Box::new(
CodexAcceptedLineFingerprintsEventRequest {
event_type: "codex_accepted_line_fingerprints",
event_params: CodexAcceptedLineFingerprintsEventParams {
event_type: "codex.accepted_line_fingerprints",
turn_id: "turn-1".to_string(),
thread_id: thread_id.to_string(),
product_surface: Some("codex".to_string()),
model_slug: Some("gpt-5.1-codex".to_string()),
completed_at: 1,
repo_hash: None,
accepted_added_lines: 1,
accepted_deleted_lines: 0,
line_fingerprints: Vec::new(),
},
},
))
}
fn sample_regular_track_event(thread_id: &str) -> TrackEventRequest {
TrackEventRequest::SkillInvocation(SkillInvocationEventRequest {
event_type: "skill_invocation",
skill_id: format!("skill-{thread_id}"),
skill_name: "doc".to_string(),
event_params: SkillInvocationEventParams {
product_client_id: None,
skill_scope: None,
plugin_id: None,
repo_url: None,
thread_id: Some(thread_id.to_string()),
turn_id: Some("turn-1".to_string()),
invoke_type: Some(InvocationType::Explicit),
model_slug: Some("gpt-5.1-codex".to_string()),
},
})
}
#[cfg(debug_assertions)]
fn unique_capture_path(name: &str) -> PathBuf {
let nonce = SystemTime::now()
.duration_since(SystemTime::UNIX_EPOCH)
.expect("system clock should be after Unix epoch")
.as_nanos();
std::env::temp_dir().join(format!(
"codex-analytics-{name}-{}-{nonce}.jsonl",
std::process::id()
))
}
fn client_with_receiver() -> (AnalyticsEventsClient, mpsc::Receiver<AnalyticsFact>) {
let (sender, receiver) = mpsc::channel(8);
let queue = AnalyticsEventsQueue {
sender,
app_used_emitted_keys: Arc::new(Mutex::new(HashSet::new())),
plugin_used_emitted_keys: Arc::new(Mutex::new(HashSet::new())),
};
(AnalyticsEventsClient { queue: Some(queue) }, receiver)
}
#[test]
#[cfg(debug_assertions)]
fn analytics_destination_uses_explicit_capture_file() {
let capture_path = unique_capture_path("destination");
let destination = AnalyticsEventsDestination::from_base_url_and_capture_file(
"https://chatgpt.com/backend-api/".to_string(),
Some(capture_path.clone()),
);
assert_eq!(
destination,
AnalyticsEventsDestination::CaptureFile {
path: capture_path.clone()
}
);
assert_eq!(
fs::read_to_string(&capture_path).expect("read capture file"),
""
);
#[cfg(unix)]
{
use std::os::unix::fs::PermissionsExt;
let mode = fs::metadata(&capture_path)
.expect("read capture file metadata")
.permissions()
.mode();
assert_eq!(mode & 0o777, 0o600);
}
fs::remove_file(capture_path).expect("remove capture file");
}
#[test]
fn analytics_destination_uses_http_without_capture_file() {
let destination = AnalyticsEventsDestination::from_base_url_and_capture_file(
"https://chatgpt.com/backend-api/".to_string(),
/*capture_file*/ None,
);
assert_eq!(
destination,
AnalyticsEventsDestination::Http {
url: "https://chatgpt.com/backend-api/codex/analytics-events/events".to_string()
}
);
}
#[test]
#[cfg(not(debug_assertions))]
fn analytics_destination_ignores_capture_file_in_release() {
let destination = AnalyticsEventsDestination::from_base_url_and_capture_file(
"https://chatgpt.com/backend-api/".to_string(),
Some(std::path::PathBuf::from("ignored.jsonl")),
);
assert_eq!(
destination,
AnalyticsEventsDestination::Http {
url: "https://chatgpt.com/backend-api/codex/analytics-events/events".to_string()
}
);
}
#[tokio::test]
#[cfg(debug_assertions)]
async fn capture_file_writes_exact_serialized_request() {
let capture_path = unique_capture_path("single");
let destination = AnalyticsEventsDestination::CaptureFile {
path: capture_path.clone(),
};
let event = sample_regular_track_event("thread-1");
let expected_event = serde_json::to_value(&event).expect("serialize expected event");
let auth = codex_login::CodexAuth::create_dummy_chatgpt_auth_for_testing();
send_track_events_request(&auth, &destination, vec![event]).await;
let contents = fs::read_to_string(&capture_path).expect("read capture file");
let lines = contents.lines().collect::<Vec<_>>();
assert_eq!(lines.len(), 1);
let payload: serde_json::Value =
serde_json::from_str(lines[0]).expect("parse captured payload");
assert_eq!(payload, serde_json::json!({"events": [expected_event]}));
fs::remove_file(capture_path).expect("remove capture file");
}
#[tokio::test]
#[cfg(debug_assertions)]
async fn capture_file_writes_final_batches_as_separate_lines() {
let capture_path = unique_capture_path("batches");
let destination = AnalyticsEventsDestination::CaptureFile {
path: capture_path.clone(),
};
let auth = codex_login::CodexAuth::create_dummy_chatgpt_auth_for_testing();
let events = vec![
sample_regular_track_event("thread-1"),
sample_accepted_line_fingerprint_event("thread-2"),
sample_regular_track_event("thread-3"),
];
for batch in track_event_request_batches(events) {
send_track_events_request(&auth, &destination, batch).await;
}
let contents = fs::read_to_string(&capture_path).expect("read capture file");
let payloads = contents
.lines()
.map(|line| serde_json::from_str::<serde_json::Value>(line).expect("parse capture line"))
.collect::<Vec<_>>();
assert_eq!(payloads.len(), 3);
assert_eq!(payloads[0]["events"][0]["skill_id"], "skill-thread-1");
assert_eq!(
payloads[1]["events"][0]["event_type"],
"codex_accepted_line_fingerprints"
);
assert_eq!(payloads[2]["events"][0]["skill_id"], "skill-thread-3");
fs::remove_file(capture_path).expect("remove capture file");
}
#[test]
#[cfg(debug_assertions)]
fn capture_write_failure_still_consumes_delivery() {
let capture_path = unique_capture_path("missing-parent").join("events.jsonl");
let destination = AnalyticsEventsDestination::CaptureFile { path: capture_path };
let payload = crate::events::TrackEventsRequest {
events: vec![sample_regular_track_event("thread-1")],
};
assert!(capture_track_events_request(&destination, &payload));
}
fn sample_turn_start_request() -> ClientRequest {
ClientRequest::TurnStart {
request_id: RequestId::Integer(1),
params: TurnStartParams {
thread_id: "thread-1".to_string(),
client_user_message_id: None,
input: Vec::new(),
..Default::default()
},
}
}
fn sample_turn_steer_request() -> ClientRequest {
ClientRequest::TurnSteer {
request_id: RequestId::Integer(2),
params: TurnSteerParams {
thread_id: "thread-1".to_string(),
expected_turn_id: "turn-1".to_string(),
client_user_message_id: None,
input: Vec::new(),
responsesapi_client_metadata: None,
additional_context: None,
},
}
}
fn sample_thread_archive_request() -> ClientRequest {
ClientRequest::ThreadArchive {
request_id: RequestId::Integer(3),
params: ThreadArchiveParams {
thread_id: "thread-1".to_string(),
},
}
}
fn sample_thread(thread_id: &str) -> Thread {
Thread {
id: thread_id.to_string(),
session_id: format!("session-{thread_id}"),
forked_from_id: None,
parent_thread_id: None,
preview: "first prompt".to_string(),
ephemeral: false,
model_provider: "openai".to_string(),
created_at: 1,
updated_at: 2,
recency_at: Some(2),
status: AppServerThreadStatus::Idle,
path: None,
cwd: test_path_buf("/tmp").abs(),
cli_version: "0.0.0".to_string(),
source: AppServerSessionSource::Exec,
thread_source: None,
agent_nickname: None,
agent_role: None,
git_info: None,
name: None,
turns: Vec::new(),
}
}
fn sample_thread_start_response() -> ClientResponsePayload {
ClientResponsePayload::ThreadStart(ThreadStartResponse {
thread: sample_thread("thread-1"),
model: "gpt-5".to_string(),
model_provider: "openai".to_string(),
service_tier: None,
cwd: test_path_buf("/tmp").abs(),
runtime_workspace_roots: Vec::new(),
instruction_sources: Vec::new(),
approval_policy: AppServerAskForApproval::OnFailure,
approvals_reviewer: AppServerApprovalsReviewer::User,
sandbox: AppServerSandboxPolicy::DangerFullAccess,
active_permission_profile: None,
reasoning_effort: None,
multi_agent_mode: Default::default(),
})
}
fn sample_thread_resume_response() -> ClientResponsePayload {
ClientResponsePayload::ThreadResume(ThreadResumeResponse {
thread: sample_thread("thread-2"),
model: "gpt-5".to_string(),
model_provider: "openai".to_string(),
service_tier: None,
cwd: test_path_buf("/tmp").abs(),
runtime_workspace_roots: Vec::new(),
instruction_sources: Vec::new(),
approval_policy: AppServerAskForApproval::OnFailure,
approvals_reviewer: AppServerApprovalsReviewer::User,
sandbox: AppServerSandboxPolicy::DangerFullAccess,
active_permission_profile: None,
reasoning_effort: None,
multi_agent_mode: Default::default(),
initial_turns_page: None,
})
}
fn sample_thread_fork_response() -> ClientResponsePayload {
ClientResponsePayload::ThreadFork(ThreadForkResponse {
thread: sample_thread("thread-3"),
model: "gpt-5".to_string(),
model_provider: "openai".to_string(),
service_tier: None,
cwd: test_path_buf("/tmp").abs(),
runtime_workspace_roots: Vec::new(),
instruction_sources: Vec::new(),
approval_policy: AppServerAskForApproval::OnFailure,
approvals_reviewer: AppServerApprovalsReviewer::User,
sandbox: AppServerSandboxPolicy::DangerFullAccess,
active_permission_profile: None,
reasoning_effort: None,
multi_agent_mode: Default::default(),
})
}
fn sample_turn_start_response() -> ClientResponsePayload {
ClientResponsePayload::TurnStart(TurnStartResponse {
turn: Turn {
id: "turn-1".to_string(),
items_view: codex_app_server_protocol::TurnItemsView::Full,
items: Vec::new(),
status: AppServerTurnStatus::InProgress,
error: None,
started_at: None,
completed_at: None,
duration_ms: None,
},
})
}
fn sample_turn_steer_response() -> ClientResponsePayload {
ClientResponsePayload::TurnSteer(TurnSteerResponse {
turn_id: "turn-2".to_string(),
})
}
#[test]
fn track_request_only_enqueues_analytics_relevant_requests() {
let (client, mut receiver) = client_with_receiver();
for (request_id, request) in [
(RequestId::Integer(1), sample_turn_start_request()),
(RequestId::Integer(2), sample_turn_steer_request()),
] {
client.track_request(/*connection_id*/ 7, request_id, &request);
assert!(matches!(
receiver.try_recv(),
Ok(AnalyticsFact::ClientRequest { .. })
));
}
let ignored_request = sample_thread_archive_request();
client.track_request(
/*connection_id*/ 7,
RequestId::Integer(3),
&ignored_request,
);
assert!(matches!(receiver.try_recv(), Err(TryRecvError::Empty)));
}
#[test]
fn track_response_only_enqueues_analytics_relevant_responses() {
let (client, mut receiver) = client_with_receiver();
for (request_id, response) in [
(RequestId::Integer(1), sample_thread_start_response()),
(RequestId::Integer(2), sample_thread_resume_response()),
(RequestId::Integer(3), sample_thread_fork_response()),
(RequestId::Integer(4), sample_turn_start_response()),
(RequestId::Integer(5), sample_turn_steer_response()),
] {
client.track_response(/*connection_id*/ 7, request_id, response);
assert!(matches!(
receiver.try_recv(),
Ok(AnalyticsFact::ClientResponse { .. })
));
}
client.track_response(
/*connection_id*/ 7,
RequestId::Integer(6),
ClientResponsePayload::ThreadArchive(ThreadArchiveResponse {}),
);
assert!(matches!(receiver.try_recv(), Err(TryRecvError::Empty)));
}
#[test]
fn track_event_request_batches_only_isolates_accepted_line_fingerprint_events() {
let batches = track_event_request_batches(vec![
sample_regular_track_event("thread-1"),
sample_regular_track_event("thread-2"),
sample_accepted_line_fingerprint_event("thread-3"),
sample_accepted_line_fingerprint_event("thread-4"),
sample_regular_track_event("thread-5"),
sample_regular_track_event("thread-6"),
]);
assert_eq!(batches.len(), 4);
assert_eq!(batches[0].len(), 2);
assert_eq!(batches[1].len(), 1);
assert_eq!(batches[2].len(), 1);
assert_eq!(batches[3].len(), 2);
assert!(batches[1][0].should_send_in_isolated_request());
assert!(batches[2][0].should_send_in_isolated_request());
}