Skip to content

The open-source app everyone uses to manage agents at work

fix(test): gate interaction-continuation retry barrier on terminal write (#10049)

Nicky Leach39b94d0e7ab2
diff --git a/server/src/__tests__/heartbeat-process-recovery.test.ts b/server/src/__tests__/heartbeat-process-recovery.test.ts
index 1432021f91..49008825b4 100644
--- a/server/src/__tests__/heartbeat-process-recovery.test.ts
+++ b/server/src/__tests__/heartbeat-process-recovery.test.ts
@@ -2211,7 +2211,21 @@ describeEmbeddedPostgres("heartbeat orphaned process recovery", () => {
.select()
.from(heartbeatRuns)
.where(eq(heartbeatRuns.agentId, agentId));
- return rows.length >= 2 ? rows : null;
+ if (rows.length < 2) return null;
+ // Gate on the *terminal* write of the background recovery, not the
+ // intermediate retry-run commit. recordPlanApprovalResumeFailureRetry
+ // writes the system comment first and updates the interaction
+ // result.resumeFailure last (heartbeat.ts:5627 then :5632), so once
+ // resumeFailure.status is observed the comment + issue update are also
+ // committed and every assertion below is race-free.
+ const interactionRow = await db
+ .select({ result: issueThreadInteractions.result })
+ .from(issueThreadInteractions)
+ .where(eq(issueThreadInteractions.id, interactionId))
+ .then((interactionRows) => interactionRows[0] ?? null);
+ const result = interactionRow?.result ?? null;
+ const resumeFailure = result && "resumeFailure" in result ? result.resumeFailure : null;
+ return resumeFailure?.status === "retrying" ? rows : null;
});
expect(runs).toHaveLength(2);