feat: Rework CancelReason

[kalyan02]

Summary

---

Note

Low Risk
Additive enum variants with docs only; low risk unless consumers use exhaustive matches or strict schema versioning without handling new snake_case values.

Overview
Extends the public CancelReason enum with five new matching-engine / lifecycle variants so CancelOrderV1, CancelTriggerOrderV1, and CancelTwapV1 can report more specific cancellation causes.

TwapExecutionFailed mirrors TriggerExecutionFailed: a TWAP slice fired but could not execute (runtime error, no liquidity, etc.), with detail expected on companion reject events. AutoDeleverage covers resting orders cleared when a position is force-closed via ADL. Expired is for makers removed during matching after their expiry time. TwapInsufficientMargin and MakerInsufficientMargin distinguish post-trade margin failures where the fill is rolled back and the TWAP or maker order is cancelled.

No other files change in this diff; wiring to emit these reasons would live elsewhere.

^{Reviewed by Cursor Bugbot for commit 873e157. Bugbot is set up for automated code reviews on this repo. Configure here.}

[cubic-dev-ai]

No issues found across 1 file

[cursor]

Cursor Bugbot has reviewed your changes and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit 342a3d9. Configure here.}

[cubic-dev-ai]

1 issue found across 1 file (changes from recent commits).

<sub>Tip: Review your code locally with the cubic CLI to iterate faster.

Re-trigger cubic</sub>

[cubic-dev-ai]

1 issue found across 1 file (changes from recent commits).

<sub>Tip: Review your code locally with the cubic CLI to iterate faster.

Re-trigger cubic</sub>

[cubic-dev-ai]

1 issue found across 1 file (changes from recent commits).

Prompt for AI agents (unresolved issues)

Check if these issues are valid — if so, understand the root cause of each and fix them. If appropriate, use sub-agents to investigate and fix each issue separately.


<file name="src/event.rs">

<violation number="1" location="src/event.rs:727">
P1: `CancelOrderV1.order_id` was changed from the plain type `OrderId` to the enum `OrderReference`, which changes the serialized wire format for an existing public event variant. Because `Event` derives both `borsh` and `serde`, downstream consumers decoding `CancelOrderV1` will see a different binary layout (discriminant + payload) and a different JSON shape (tagged enum object instead of a scalar). If this variant is already emitted anywhere (unlike the removed `CancelTriggerOrderV1`/`CancelTwapV1`, which were reserved because they were never deployed), existing indexers, clients, and generated schemas can fail to deserialize or silently misinterpret the field. Consider either retaining backward compatibility—e.g., by keeping separate V1 variants for each order type—or ensuring all downstream consumers are version-locked and migrated together.</violation>
</file>

<sub>Tip: Review your code locally with the cubic CLI to iterate faster.

Re-trigger cubic</sub>

[cubic-dev-ai]

P1: CancelOrderV1.order_id was changed from the plain type OrderId to the enum OrderReference, which changes the serialized wire format for an existing public event variant. Because Event derives both borsh and serde, downstream consumers decoding CancelOrderV1 will see a different binary layout (discriminant + payload) and a different JSON shape (tagged enum object instead of a scalar). If this variant is already emitted anywhere (unlike the removed CancelTriggerOrderV1/CancelTwapV1, which were reserved because they were never deployed), existing indexers, clients, and generated schemas can fail to deserialize or silently misinterpret the field. Consider either retaining backward compatibility—e.g., by keeping separate V1 variants for each order type—or ensuring all downstream consumers are version-locked and migrated together.

Prompt for AI agents

Check if this issue is valid — if so, understand the root cause and fix it. At src/event.rs, line 727:

<comment>`CancelOrderV1.order_id` was changed from the plain type `OrderId` to the enum `OrderReference`, which changes the serialized wire format for an existing public event variant. Because `Event` derives both `borsh` and `serde`, downstream consumers decoding `CancelOrderV1` will see a different binary layout (discriminant + payload) and a different JSON shape (tagged enum object instead of a scalar). If this variant is already emitted anywhere (unlike the removed `CancelTriggerOrderV1`/`CancelTwapV1`, which were reserved because they were never deployed), existing indexers, clients, and generated schemas can fail to deserialize or silently misinterpret the field. Consider either retaining backward compatibility—e.g., by keeping separate V1 variants for each order type—or ensuring all downstream consumers are version-locked and migrated together.</comment>

<file context>
@@ -732,28 +724,16 @@ pub enum Event<Address> {
     CancelOrderV1 {
         user_address: Address,
-        order_id: OrderId,
+        order_id: OrderReference,
         market_id: MarketId,
         execution_timestamp: UnixTimestampMicros,
</file context>

[vmmon]

CancelOrderV1 was never deployed either. So no breakage here.

[cubic-dev-ai]

You're right — if CancelOrderV1 was never deployed, then there’s no downstream breakage from changing its wire format here.