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

Timeline: draw earlier retry attempts on the thread that ran them (retryAttempts[].threadId)

Aperta
#334 0 commenti 0 reazioni 0 assegnatari Vedi su GitHub

Nessuno ha ancora preso questa issue.

Valutazione

Difficoltà
4/5
Tempo stimato
3-5 giorni
Idoneità per principianti
55/100
Tipo di issue
Funzionalità
Chiarezza
Specificata chiaramente
Stato di attività
Attiva
Stack tecnologico
node.js, typescript

Direzione di ricerca

Start with src/ctrf/timeline/collect.ts, especially readTests, arrange() and placeAttempts(), then review the timeline notes in src/ctrf/timeline/report.ts. Check the timeline tests and sample at ctrf-reports/timeline/ctrf-report.json; the issue does not name specific test files. Done when attempts with recorded threadId appear on their recorded worker, missing IDs retain current placement, and the docs and sample reflect the change.

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

Descrizione

Summary

CTRF 0.3.0 adds an optional threadId to attempt history entries (retryAttempts[].threadId), so each earlier attempt can record the thread or worker that ran it. The Test-level threadId now explicitly describes the final attempt. See ctrf-io/ctrf#74 and ctrf-io/ctrf#86, part of the 0.3.0 release tracked in ctrf-io/ctrf#75.

The timeline report (#279) should use it to draw each earlier attempt on the thread that actually ran it, and keep today's placement only for attempts that don't record a thread.

How it works today

Earlier attempts don't record a thread, so readTests gives every attempt its test's threadId, which is the thread of the final attempt (collect.ts#L353-L372). arrange() packs each thread's final results, and placeAttempts() tries to fit each earlier attempt into a gap on the first four lanes of that thread (collect.ts#L908, collect.ts#L1019). Attempts that fit nowhere go to a separate "Runs before a retry" lane, with a note that such runs do not record where they ran (report.ts#L748-L751).

When a retry runs on a different worker than the failed attempt, that guess is wrong in one of two ways:

  • The attempt fits a gap on the final attempt's thread. It is drawn on a worker that never ran it, and nothing in the chart says so.
  • It doesn't fit. It moves to the "Runs before a retry" lane, and the worker that actually ran it shows an idle gap at that time.

Either way, the tests drawn next to the failed attempt are the wrong neighbours. That matters most when looking for order-dependent or shared-state flakiness.

Proposed change

  1. Read the field. In readTests, read attempt.threadId with the same normalization as the test's (normalizeThreadId) and keep it on the attempt. An attempt without one inherits its test's thread, as today, and is marked as inherited.
  2. Place recorded attempts like final results. In arrange(), put attempts with a recorded thread into that thread's group and pack them together with its final results. Only inherited attempts go through placeAttempts() and can end up in "Runs before a retry".
  3. Keep overlap handling. The spec doesn't guarantee that a worker runs one test at a time, since some frameworks run tests concurrently within a worker. A recorded attempt that overlaps other tests on its thread splits the thread into extra lanes, as final results already do, instead of being moved out.
  4. Update the counts and notes. The stray-attempt note counts only attempts without a recorded thread, for which its wording stays true. Thread counts and the maxLanes check include threads that appear only in attempt history.
  5. Types. The ctrf package (0.5.0) types AttemptHistoryEntry without threadId and validates against spec 0.2.0. Until it ships spec 0.3.0, read the field through a narrow cast; then bump ctrf and drop the cast. Merging already keeps the field, because mergeReports copies tests as they are, but it can't validate such reports against 0.2.0, so the merged specVersion falls back to the inputs' version.
  6. Docs and sample. Mention attempt threadId in the timeline section of docs/report-showcase.md, and give the retried test in ctrf-reports/timeline/ctrf-report.json an attempt on another worker.

Reports without attempt threadId render exactly as they do now.

How it improves the timeline

  • Each earlier attempt is drawn on the worker that ran it, so per-worker busy and idle time is accurate.
  • A failed attempt is shown next to the tests that really ran beside it on its worker. That is where order-dependent and shared-state flakiness shows up.
  • A retry that moved to another worker is visible as such: the failed attempt on one row, the final attempt on another.
  • A worker whose only results are failed attempts, such as one replaced after a failure, gets its own row. Today it doesn't appear.
  • The "Runs before a retry" lane and its note remain only for reports that don't record attempt threads.

Dependencies

  • ctrf-io/ctrf#86: the spec change, in the 0.3.0 stack (ctrf-io/ctrf#75).
  • A ctrf npm release with spec 0.3.0, for the types.
  • Reporters that fill in retryAttempts[].threadId. Until they do, the timeline stays as it is.
Lingua principale
TypeScript
Stelle
380
Fork
43
Merge medio
3g 11h
PR unite (30g)
2

Preparare l'ambiente

Apri in Codespaces

Avvia il container di sviluppo del progetto nel browser, con il tuo account GitHub.

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 ctrf-io/github-test-reporter

Tutte le issue di ctrf-io/github-test-reporter

Issue simili

Altre issue su TypeScript

Ricevi le nuove issue nella tua casella

Un breve riepilogo di issue GitHub adatte ai principianti.