[Krrishkumar125]

Fixes #635.

Since 0.17.0, buildRunner turns a worker pool (or cron) rejection into a stop(...) call and then .catch(noop)s it, so Promise.all([cp, wp]) always resolves. runner.promise therefore resolves after a fatal pool error, and the CLI exits 0.

Change: the exit handlers still call stop() with the same log reason, but on the error path they re-throw the original error (and only swallow a failure of stop() itself). The existing rejection branch of Promise.all then runs and runner.promise rejects with the pool's error. The clean-exit path is unchanged.

Tests (__tests__/runner.promise.test.ts):

• the reproduction from the issue (a trigger refuses job completion, completeJobBatchDelay: 0): fails on main with "resolved", passes with this change
• runner.stop() still resolves runner.promise

Full suite passes locally on PostgreSQL 16 (24 suites, 107 tests), plus yarn lint (no new warnings) and yarn prepack. Added a patch changeset.

@priyansh00soni asked to take this on Oct 3; I didn't see a PR since, but happy to close this if you already have one in progress.

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.