Hacktoberfest 2026: die Issues, die Maintainer für den Oktober markiert haben – offen und einsteigerfreundlich. Hacktoberfest-Issues durchsuchen

CI deploy fails with "Cloudflare session has no bearer for version-aware upload" (runner-only; passed 10h earlier with the same token)

Offen
#627 0 Kommentare 0 Reaktionen 0 zugewiesene Personen Auf GitHub ansehen

Maintainer antworten meist innerhalb von 1 Tag

Dieses Issue hat noch niemand übernommen.

Bewertung

Schwierigkeit
4/5
Geschätzter Aufwand
3-5 Tage
Anfängerfreundlichkeit
48/100
Issue-Typ
Bug
Klarheit
Größtenteils klar
Aktivitätsstatus
Aktiv
Tech-Stack
bun, github-actions, typescript

Rechercherichtung

Start by reading the Cloudflare deploy transport in deploy-transport-JCxw7QDi.mjs, especially deployChildEnv and runWranglerAuthToken, then trace the upload and trigger-sync recovery paths. Compare behavior with and without CLOUDFLARE_API_TOKEN in a fresh checkout and inspect how version bindings and secrets are preserved. Done means CI can upload with the environment token and recovery does not activate a version missing the previous bindings.

Vom Indexierungsmodell aus dem Issue-Text verfasst.

Beschreibung

What happened

void deploy --platform cloudflare (void 1.3.0, GitHub Actions, ubuntu-latest, bun 1.4.3) fails at the very start of the upload step:

◇ Checking the current Worker deployment...   (12 s)
◇ Uploading the Worker and assets...          (0.2 s)
■ Worker deployment stopped.
deploy: Cloudflare Worker Version upload failed (exit 1); production traffic is unchanged.
Cloudflare session has no bearer for version-aware upload.

The same workflow uploaded and activated successfully ~13 hours earlier (2026-10-27 12:52 UTC) with the same CLOUDFLARE_API_TOKEN secret, same repo config, same workflow file. Between the two runs nothing on the repo/auth side changed; the second failure then reproduced identically on three consecutive runs.

Repro status
  • Not reproducible locally: a fresh git archive of the same commit (no .void/ state, i.e. exactly the CI checkout), bun install, then CLOUDFLARE_API_TOKEN=<same shape: user token, Edit Cloudflare Workers template> bunx void deploy --platform cloudflare succeeds on macOS (pinned upload path included — usesCf is structural, so runWranglerAuthToken resolved the api_token fine).
  • On the runner the same getToken() (runWranglerAuthToken → wrangler auth.token under deployChildEnv) returns null.
Suspected cause

deployChildEnv hardcodes CLOUDFLARE_AUTH_USE_KEYRING = "true" (deploy-transport-JCxw7QDi.mjs). On a headless GitHub runner there is no user keyring session; if wrangler's keyring init fails/hangs there, auth.token exits non-clean and runWranglerAuthToken returns null → "no bearer". On macOS (local repro) the keyring exists, so the env-token path proceeds. Consistent with "worked at 12:52, broke later with zero auth-side change" if a runner image roll changed the keyring situation.

Request: when CLOUDFLARE_API_TOKEN is present in the environment, either don't force the keyring, or fall back from the pinned uploader to the plain wrangler upload instead of failing.

Second issue hit on the same integration

The earlier 12:52 run got past upload and activation but failed at trigger synchronization; the recovery left an empty, binding-less version active (no main, no assets, compat date defaulted to 2026-10-19), which replaced our production worker and wiped both Worker secrets until we redeployed manually. void's upload patch sets keep_vars: true but never keep_secrets, and the fresh-checkout CI path has no .void/ link state. Two suggestions: (1) the trigger-sync recovery should re-activate the previous version's full bindings, not an empty shape; (2) uploads from a fresh checkout should preserve secret_text bindings the way they preserve vars.

Happy to provide the JSONL logs (sanitized) — the top-level events only show deploy_start → deploy_target → cloudflare_deploy_start → cloudflare_deploy_failed; no per-child detail is mirrored even with --debug.

Vorherrschende Sprache
TypeScript
Sterne
580
Forks
56
Ø Merge
11 Std. 32 Min.
Gemergte PRs (30 T.)
158

Entwicklungsumgebung

Dieses Projekt bietet weder Dev-Container noch Dockerfile noch Beitragsleitfaden – die Einrichtung liegt bei Ihnen. Beginnen Sie mit der README; die allgemeinen Schritte stehen in unserem Leitfaden für den ersten Beitrag.

Erste Schritte

  1. Lesen Sie das ganze Issue und danach den Beitragsleitfaden des Projekts.
  2. Schreiben Sie ins Issue, dass Sie es übernehmen — das erspart doppelte Arbeit.
  3. Forken Sie das Repository und arbeiten Sie in einem Branch.
  4. Öffnen Sie einen Pull Request, der die Issue-Nummer nennt.

Mehr aus voidzero-dev/void

Alle Issues in voidzero-dev/void

Ähnliche Issues

Weitere Issues zu TypeScript

Neue Issues direkt in Ihr Postfach

Eine kurze Übersicht über anfängerfreundliche GitHub-Issues.