Skip to content

fix(chat): keep context-limit failures in the pending queue #8164

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
loopx-agent merged 1 commit into from
Nov 6, 2026
Merged

fix(chat): keep context-limit failures in the pending queue #8164

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
7 changes: 4 additions & 3 deletions loopx/control_plane/collaboration/conversation_binding.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -509,7 +509,7 @@ export function planBoundConversationRequest(params: JsonObject): JsonObject {
return {pending: !returned && row.command === null && row.status === "accepted"
&& turn?.session_id === row.session_id && turn?.turn_id === row.turn_id
&& (expandedResult || metadataOwed || answerReviewOwed || unfinished || turn?.status === "failed"
&& ["usage_limit_exceeded", "response_timeout", "upstream_transport_failure",
&& ["usage_limit_exceeded", "response_timeout", "upstream_transport_failure", "context_window_exceeded",
"antigravity_rpc_error", "antigravity_transport_error", "antigravity_step_error",
"manager_authorization_unavailable", "runtime_error", "server_restarted"].includes(String(turn.error_code)))
&& !(current?.session_id === row.session_id && current?.status === "closed")
Expand Down Expand Up @@ -1141,7 +1141,8 @@ function planExhaustedStewardAnswer(recovery: JsonObject, row: JsonObject, curre
&& Number(item.manager_context_version) <= version)
.map(item => sessionIdentity(item.thread_id, "original native thread"))];
const priorNative = budget.native_observation as JsonObject | undefined;
const nativeThreads = [...new Set([...(threads.includes(priorNative?.thread_id) ? [priorNative!.thread_id] : []), ...threads])];
const nativeThreads = [...new Set([...(typeof priorNative?.thread_id === "string" && threads.includes(priorNative.thread_id)
? [priorNative.thread_id] : []), ...threads])];
const candidate = {...no, recovery_candidate: true, answer_review: true, read_only_review: true,
client_turn_id: client, session_id: current.session_id, failed_turn_id: failed.turn_id,
native_thread_ids: nativeThreads};
Expand All @@ -1158,7 +1159,7 @@ function planExhaustedStewardAnswer(recovery: JsonObject, row: JsonObject, curre
|| !Number.isFinite(now) || !Number.isFinite(ended) || now < ended + 60_000) return no;
if (recovery.native_observation == null) return candidate;
const proof = requireJsonObject(recovery.native_observation, "exhausted original request inventory");
if (!nativeThreads.includes(proof.thread_id) || proof.turn_id !== failed.upstream_turn_id
if (typeof proof.thread_id !== "string" || !nativeThreads.includes(proof.thread_id) || proof.turn_id !== failed.upstream_turn_id
|| proof.only_original_user_item !== true || proof.codex_error_info !== "usageLimitExceeded") return candidate;
const capacity = nativeRecoveryCapacity(recovery, row, current, candidate);
return {...capacity, recovery_eligible: capacity.recovery_eligible === true && capacity.answer_executor === "codex"};
Expand Down
5 changes: 3 additions & 2 deletions tests/control_plane_ts/conversation_binding.test.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -68,9 +68,10 @@ test("exhausted business quota permits one read-only answer, never a fresh busin
{executor_configuration: {...configuration, configuration_revision: "changed"}},
{metadata_capacity: {alternate_executor: {...capacity.alternate_executor, observed_at: at - 121}}},
{metadata_capacity: {alternate_executor: {...capacity.alternate_executor, observed_at: at + 1}}},
...[{thread_id: "foreign"}, {turn_id: "foreign"}, {only_original_user_item: false},
...[{thread_id: "foreign"}, {thread_id: null}, {thread_id: 7}, {turn_id: "foreign"}, {only_original_user_item: false},
{codex_error_info: "other"}].map(value => ({native_observation: {...proof, ...value}})),
...[{delta_count: 1}, {response: {}}, {status: "running"}, {message: "changed"},
{error_code: "context_window_exceeded"},
{client_turn_id: "uncorrelated"}].map(value => ({failed_turn: {...failed, ...value}}))]) {
assert.notEqual(planBoundConversationRequest({...packet,
context_recovery: {...packet.context_recovery, ...change}}).recovery_eligible, true);
Expand Down Expand Up @@ -1447,7 +1448,7 @@ test("quota backlog projection retires explicit lifecycle stops, not unavailable
}
});

for (const errorCode of ["response_timeout", "upstream_transport_failure",
for (const errorCode of ["response_timeout", "upstream_transport_failure", "context_window_exceeded",
"antigravity_rpc_error", "antigravity_transport_error", "antigravity_step_error", "runtime_error", "server_restarted"]) {
test(`delivered ${errorCode} notices retain the exact failed request, including other executors`, () => {
const request = {request_ref: "f".repeat(24), command: null, status: "accepted",
Expand Down
2 changes: 1 addition & 1 deletion tests/test_lark_private_conversations.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1325,7 +1325,7 @@ def test_delivered_quota_failure_stays_pending_across_restart_without_replaying(


@pytest.mark.parametrize("command", ["stop", "new"])
@pytest.mark.parametrize("error_code", ["response_timeout", "upstream_transport_failure"])
@pytest.mark.parametrize("error_code", ["response_timeout", "upstream_transport_failure", "context_window_exceeded"])
def test_delivered_timeout_stays_pending_without_replaying_and_honors_stop(ordinary, monkeypatch, command, error_code): # noqa: F811
from loopx.chat_agent import CodexChatAgentError
from loopx.capabilities.native_chat.codex_adapter import CodexAppServerAdapter
Expand Down
Loading