OrderDeletedError

Extends Error

CONFIRMED order-not-found — "the exchange definitively reports there is no order under this id anymore" (e.g. the user cancelled it manually on the exchange, or it was liquidated / garbage-collected externally).

Where to throw it

Only from the ORDER CHECKS (the onOrderCheck ping channel), i.e. any of:

• Broker.useBrokerAdapter → onOrderActiveCheck / onOrderScheduleCheck;
• action schema callbacks.onOrderCheck / handler IAction.orderCheck (deprecated channel);
• a listenCheck listener.

What the framework does

The throw propagates UNWRAPPED through every layer (the runtime brand survives) and resolves into the IBrokerOrderVerdict { reason: "deleted" } — terminal IMMEDIATELY, bypassing the CC_ORDER_CHECK_RETRY_ATTEMPTS tolerance counter:

• event.type === "active" (open position): the position is closed with closeReason "closed" WITHOUT re-confirmation through the close gate — the ping already established the order is gone, re-asking the broker would be redundant.
• event.type === "schedule" (resting entry order): the scheduled signal is cancelled with reason "user". The schedule-cancelled lifecycle event still reaches the broker adapter (onSignalScheduleCancelled); cancelling an already-gone order there is a no-op.

Loudness: errorEmitter fires. exitEmitter does NOT fire — a confirmed not-found is a business fact about ONE order, not a fatal network condition. Compare with transient-failure exhaustion of the same check, which DOES signal a fatal exit.

When it is appropriate

Throw ONLY on the exchange's definitive "order not found by event.signalId" response. Two critical distinctions:

• A FILLED resting order is NOT a deleted order. If the scheduled entry actually filled, confirm it via commitActivateScheduled instead — a throw here is a terminal CANCEL, not an activation. Same for an open position whose TP/SL order filled on the exchange: report it via commitCreateTakeProfit / commitCreateStopLoss so the close carries the true closeReason, instead of collapsing it into a generic "closed".
• Network trouble is NOT a deleted order. On timeout / 5xx / rate limit / disconnect throw a plain Error or OrderTransientError: the framework tolerates up to CC_ORDER_CHECK_RETRY_ATTEMPTS consecutive transient failures (the order is assumed still open, event.attempt increments) before acting terminally. Reporting a blip as "deleted" kills a live position on the spot.

Nuances

• Context-specific. This error belongs to the CHECKS. Throwing it from a GATE (onOrderOpenCommit / onOrderCloseCommit / callbacks.onOrderSync) is a userspace protocol violation: it is INTENTIONALLY degraded to the "transient" verdict (bounded retry) instead of being honored as terminal.
• Nominal runtime identification. Recognized by the __type__ === Symbol.for("OrderDeletedError") brand via the static guard — never by instanceof, so it survives duplicated module instances across bundles.
• Live-only. Checks never fire in backtest (there is no live exchange to query).
• A successful check resets the consecutive-failure counter (event.attempt) to 0; this error ignores the counter entirely in both directions — it neither needs prior failures nor is delayed by remaining tolerance.
• The message is optional and purely informational; routing depends only on the brand.

Constructor

constructor(message: string);
Copy

Properties

type

__type__: symbol
Copy

Runtime brand (Symbol.for — survives duplicated module instances)

Methods

isOrderDeletedError

static isOrderDeletedError(error: object): boolean;
Copy

Nominal type guard by the runtime brand. Use this instead of instanceof: the check is based on Symbol.for, so it recognizes instances created by a DIFFERENT copy of this module (duplicated bundles, linked packages).

fromError

static fromError(error: object): OrderDeletedError;
Copy

Nominal constructor for a new OrderDeletedError from any thrown object. Use this instead of instanceof to recognize instances created by a DIFFERENT copy of this module (duplicated bundles, linked packages).