Hacktoberfest 2026: le issue che i maintainer hanno segnato per ottobre, aperte e adatte ai principianti. Sfoglia le issue Hacktoberfest

[test] Make the Codex round 4 failure-boundary regression deterministic

Aperta
  • Qualcosa non va?
#178 0 commenti 0 reazioni 0 assegnatari Vedi su GitHub

I maintainer di solito rispondono entro 1 giorno

Nessuno ha ancora preso questa issue.

Valutazione

Difficoltà
3/5
Tempo stimato
1-2 giorni
Idoneità per principianti
55/100
Tipo di issue
Bug
Chiarezza
Abbastanza chiara
Stato di attività
Attiva
Stack tecnologico
typescript
Ambito
testing

Direzione di ricerca

The flaky case is round 4 failure boundaries > bounds setup and native cleanup without waiting for an unadmitted capture in packages/worker-shim/src/adapters/codex.test.ts. Read it alongside packages/worker-shim/README.md, which describes the DeepSeek held-finalization fixture that replaces timing assumptions with explicit signals. First reproduce the failure by running that single test 30 to 50 times while a parallel coverage load runs, then find the wall-clock or scheduling assumption that lets the assertion run before invalidation. Done means the test passes 30 consecutive times under load, with the assertion and the production timeouts and budgets unchanged.

Scritto dal modello di indicizzazione a partire dal testo della issue.

Descrizione

bug

Summary

Make the Codex round 4 failure-boundary regression deterministic under CI coverage load.

Category

test or CI

Area

worker-shim Codex adapter tests

Observed behavior and evidence

In one of three full pnpm -w verify:l0-l2 runs on the CI test image at main 3a54ea25 (linux/amd64, coverage enabled), worker-shim coverage failed src/adapters/codex.test.ts > round 4 failure boundaries > bounds setup and native cleanup without waiting for an unadmitted capture with AssertionError: expected "invalidate" to be called once, but got 0 times. The same test passed in the other two runs and in the regular CI runs around it. No production change was involved, so this is an intermittent fixture timing assumption, not a demonstrated adapter defect.

Impact

Severity: minor. An intermittent failure can fail an otherwise green L0-L2 gate and costs a full rerun. It does not indicate a product regression on present evidence.

Suggested fix

Reproduce the case under repeated runs with CPU contention, for example 30 to 50 repetitions of the single test beside a parallel coverage load. Identify which wall-clock or scheduling assumption lets the assertion run before the invalidation is observed. Replace polling or elapsed-time assumptions with an explicit signal from the fixture, or control the preparation timers, as the DeepSeek held-finalization fixture now does (see packages/worker-shim/README.md). Keep the production setup and cleanup bounds and the asserted invalidation unchanged; do not add retries or widen budgets to obtain green.

Acceptance criteria

  • The test passes in at least 30 consecutive repetitions under parallel load on a supported Linux host.
  • The assertion that setup and native cleanup are bounded and that invalidation happens exactly once is unchanged.
  • No production timeout, retry, or budget is changed by the fix.

References

  • packages/worker-shim/src/adapters/codex.test.ts
  • packages/worker-shim/README.md
  • docs/specs/20260529-test_strategy.md
  • Related: #139

Filed from the v0.1.0 acceptance-program inventory.

Lingua principale
TypeScript
Stelle
0
Fork
0
Merge medio
12h 53m
PR unite (30g)
13

Preparare l'ambiente

Come iniziare

  1. Leggi tutta la issue e poi la guida ai contributi del progetto.
  2. Commenta sulla issue per dire che te ne occupi tu — evita che due persone facciano lo stesso lavoro.
  3. Fai un fork del repository e lavora su un branch.
  4. Apri una pull request che faccia riferimento al numero della issue.

Altre issue di lingkaix/openkit

Tutte le issue di lingkaix/openkit

Issue simili

Altre issue su TypeScript

Ricevi le nuove issue nella tua casella

Un breve riepilogo di issue GitHub adatte ai principianti.