fix(scheduler): add require.main guard so smoke-test require() doesn't run scheduling#241
Merged
therealbrad merged 4 commits intomainfrom Apr 24, 2026
Merged
fix(scheduler): add require.main guard so smoke-test require() doesn't run scheduling#241therealbrad merged 4 commits intomainfrom
therealbrad merged 4 commits intomainfrom
Conversation
…start workers
The previous guard combined an ESM-style import.meta check with a CJS
fallback:
if (
(typeof import.meta !== "undefined" &&
import.meta.url === pathToFileURL(process.argv[1]).href) ||
typeof import.meta === "undefined" ||
(import.meta as any).url === undefined
) { startWorker()... }
esbuild compiles each worker to CommonJS (platform: node, format: cjs)
and polyfills `import.meta` as a plain object whose `.url` is
`undefined`. At runtime that makes `import.meta.url === void 0` always
true, so the guard always fires — meaning `require("./forecastWorker")`
unintentionally invokes `startWorker()` and all its connection logic.
Three workers (forecastWorker, repoCacheWorker, testmoImportWorker)
call `process.exit(1)` synchronously when Valkey is unreachable, so the
CI smoke test added in #237 died mid-loop and releases blocked on the
smoke-test step.
Fix: replace with the canonical CJS pattern `require.main === module`.
esbuild preserves `require`/`module` in CJS output, and the smoke test
can now `require()` each worker without triggering startup side effects
— matching the intent stated in the original comment.
Drops the `pathToFileURL` import from each worker (was only used by
the old guard).
v0.22.9 smoke test surfaced two issues hidden behind the main-guard bug fixed in the previous commit: 1. generateFromUrlWorker.ts had no main guard at all — it called startGenerateFromUrlWorker() unconditionally at module scope, so require()'ing it in the smoke test attempted to construct a BullMQ Worker with no Valkey and crashed. Wrapped in `if (require.main === module)` to match the other 14 workers. 2. env.js validates DATABASE_URL / NEXTAUTH_SECRET / NEXTAUTH_URL at module-load time via @t3-oss/env-nextjs, so any worker whose transitive imports reach env.js (syncWorker, elasticsearchReindexWorker, copyMoveWorker, duplicateScanWorker, magicSelectWorker) threw during require with zod 'invalid_type' errors. Added dummy env shims at the top of smoke-test-workers.js using `||=` so real CI-provided values still win. The smoke test is verifying module-graph integrity, not runtime config correctness. Together with the main-guard fix, the smoke test should now complete cleanly for every worker.
…t run scheduling
scheduler.ts called scheduleJobs() at module top-level, so the v0.22.10
smoke test's require() of dist/scheduler.js tried to connect to Valkey
and exited 1 ('Required queues are not initialized. Cannot schedule
jobs.') even with SKIP_VALKEY_CONNECTION=true (which makes queues no-op
but still marks them unavailable).
Same pattern as the workers fix in #239/#240: gate the runtime code
with `if (require.main === module)`. Production start-workers.sh
invokes scheduler directly via `tsx scheduler.ts`, so require.main IS
module at runtime and scheduleJobs() still runs.
Contributor
Author
|
🎉 This PR is included in version 0.22.11 🎉 The release is available on GitHub release Your semantic-release bot 📦🚀 |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Third and (hopefully) final smoke-test fix, following #239 and #240.
v0.22.10 smoke test passed all 15 worker entrypoints ✓ but the scheduler still exited 1:
scheduler.tscalledscheduleJobs()at module top-level, sorequire("./dist/scheduler.js")from the smoke-test script tried to connect to Valkey and bail when queues weren't available — even withSKIP_VALKEY_CONNECTION=true(which makes the connection skip cleanly but still marks queues as unavailable).Fix
Wrap the
scheduleJobs()invocation inif (require.main === module), matching the pattern already applied to all 15 workers in #239 / #240.Production start path is unchanged:
start-workers.shrunstsx scheduler.tsdirectly, sorequire.main === moduleevaluates true andscheduleJobs()still runs.Test plan
Upserted job scheduler ...entries)🤖 Generated with Claude Code