release.yml bump-scoop-manifest pushes to archived coilyco-bridge/scoop-bucket; must be coilyco-flight-deck/scoop-bucket #589

New issue

Closed

opened 2026-07-03 21:40:50 +00:00 by coilysiren · 2 comments

coilysiren commented

2026-07-03 21:40:50 +00:00

Owner

Bug

The bump-scoop-manifest job (ward#571) hardcodes BUCKET_REPO: coilyco-bridge/scoop-bucket (release.yml:614). That repo is archived - the push 403s with remote: This repo is archived ... cannot push. The active, pushable bucket is coilyco-flight-deck/scoop-bucket (verified: archived=False, default_branch=main). ward#571's engineer inspected the archived coilyco-bridge bucket and hardcoded it.

Fix

Change BUCKET_REPO (line 614) to coilyco-flight-deck/scoop-bucket. It is referenced via ${BUCKET_REPO} at lines 638/661/716/718/725/727/730, so the single change corrects clone, push, and verify. Also fix any stale coilyco-bridge/scoop-bucket (or coilysiren/scoop-bucket) references in docs/release.md and the comment at release.yml:102.

Note

SCOOP_WRITE_TOKEN must have push access to coilyco-flight-deck/scoop-bucket (Kai is setting it for flight-deck). Once this lands, re-run bump-scoop-manifest (or the next release) pushes ward.json to the correct bucket.

ward#571 (the job), ward#576 (release-publish blocker - this is the SCOOP half).

## Bug The bump-scoop-manifest job (ward#571) hardcodes `BUCKET_REPO: coilyco-bridge/scoop-bucket` (release.yml:614). **That repo is archived** - the push 403s with `remote: This repo is archived ... cannot push`. The active, pushable bucket is **`coilyco-flight-deck/scoop-bucket`** (verified: archived=False, default_branch=main). ward#571's engineer inspected the archived coilyco-bridge bucket and hardcoded it. ## Fix Change `BUCKET_REPO` (line 614) to **`coilyco-flight-deck/scoop-bucket`**. It is referenced via `${BUCKET_REPO}` at lines 638/661/716/718/725/727/730, so the single change corrects clone, push, and verify. Also fix any stale `coilyco-bridge/scoop-bucket` (or `coilysiren/scoop-bucket`) references in docs/release.md and the comment at release.yml:102. ## Note `SCOOP_WRITE_TOKEN` must have push access to `coilyco-flight-deck/scoop-bucket` (Kai is setting it for flight-deck). Once this lands, re-run bump-scoop-manifest (or the next release) pushes ward.json to the correct bucket. ## Related ward#571 (the job), ward#576 (release-publish blocker - this is the SCOOP half).

coilyco-ops commented

2026-07-03 21:41:01 +00:00

Member

🔒 Reserved by ward agent --driver claude — container engineer-claude-ward-589 on host KAI-DESKTOP-TOWER is carrying this issue (reserved 2026-07-03T21:40:56Z). Concurrent ward agent runs are blocked until it finishes or the reservation goes stale (2h0m0s TTL); --force overrides.

Do not comment on or edit this issue to steer the run while it is reserved. The engineer seeded the body once at launch and never re-reads it, so a comment or edit reaches only human readers, never the running engineer. A correction goes to a new issue, dispatched fresh — that is the only channel that reaches a run in flight. Where the forge supports it, ward locks this conversation to make that a road-block rather than a convention (ward#494).

— Claude (she/her), via ward agent

🔒 Reserved by `ward agent --driver claude` — container `engineer-claude-ward-589` on host `KAI-DESKTOP-TOWER` is carrying this issue (reserved 2026-07-03T21:40:56Z). Concurrent `ward agent` runs are blocked until it finishes or the reservation goes stale (2h0m0s TTL); `--force` overrides. **Do not comment on or edit this issue to steer the run while it is reserved.** The engineer seeded the body once at launch and never re-reads it, so a comment or edit reaches only human readers, never the running engineer. A correction goes to a **new issue, dispatched fresh** — that is the only channel that reaches a run in flight. Where the forge supports it, ward locks this conversation to make that a road-block rather than a convention (ward#494).  — Claude (she/her), via `ward agent`

coilysiren referenced this issue from a commit

2026-07-03 21:43:37 +00:00

fix(release): point bump-scoop-manifest at pushable coilyco-flight-deck/scoop-bucket (closes #589)

coilysiren closed this issue

2026-07-03 21:43:37 +00:00

coilysiren commented

2026-07-03 21:44:07 +00:00

Author

Owner

WARD-OUTCOME: done - BUCKET_REPO repointed to the pushable coilyco-flight-deck/scoop-bucket, plus stale bucket refs corrected across release.yml, docs, and upgrade.go/test.

Small, clean change that felt exactly as scoped. The single BUCKET_REPO env var flows through clone/push/verify via ${BUCKET_REPO}, so the one-line fix at line 614 does the real work. What made it slightly bigger than the issue named: grepping turned up three more stale spellings of the old bucket (a coilysiren/scoop-bucket in docs/release-binaries.md and two Go comments in upgrade.go/upgrade_test.go) - all cosmetic, but they'd have kept pointing readers at the wrong repo, so I corrected them too for a clean sweep. Left the coilysiren: username in the push URL alone: it's the auth username (Forgejo ignores it, uses the token) and matches the tap job's convention exactly.

Nothing fought back - build and tests green, pre-commit passed, fast-forward to main. High confidence in the code correctness. The one thing I can't verify from here is the runtime prerequisite the issue itself flagged: SCOOP_WRITE_TOKEN must actually carry push scope to the new bucket. That's Kai's to provision, and the next release (or a manual bump-scoop-manifest re-run) is the real end-to-end test. No follow-ups worth filing beyond that token confirmation.

WARD-OUTCOME: done - BUCKET_REPO repointed to the pushable coilyco-flight-deck/scoop-bucket, plus stale bucket refs corrected across release.yml, docs, and upgrade.go/test. Small, clean change that felt exactly as scoped. The single BUCKET_REPO env var flows through clone/push/verify via ${BUCKET_REPO}, so the one-line fix at line 614 does the real work. What made it slightly bigger than the issue named: grepping turned up three more stale spellings of the old bucket (a `coilysiren/scoop-bucket` in docs/release-binaries.md and two Go comments in upgrade.go/upgrade_test.go) - all cosmetic, but they'd have kept pointing readers at the wrong repo, so I corrected them too for a clean sweep. Left the `coilysiren:` username in the push URL alone: it's the auth username (Forgejo ignores it, uses the token) and matches the tap job's convention exactly. Nothing fought back - build and tests green, pre-commit passed, fast-forward to main. High confidence in the code correctness. The one thing I can't verify from here is the runtime prerequisite the issue itself flagged: SCOOP_WRITE_TOKEN must actually carry push scope to the new bucket. That's Kai's to provision, and the next release (or a manual bump-scoop-manifest re-run) is the real end-to-end test. No follow-ups worth filing beyond that token confirmation.