chore(workflows): self-bootstrap release.yml @v0.8.0-rc.1 (ADR-0007 dogfood) #118
No reviewers
Labels
No labels
bump
major
bump
minor
bump
patch
kind/bug
kind/chore
kind/docs
kind/feature
priority/critical
priority/high
priority/low
priority/medium
size/L
size/M
size/S
size/XL
No milestone
No project
No assignees
2 participants
Notifications
Due date
No due date set.
Dependencies
No dependencies set.
Reference
frankenbit/release-toolkit!118
Loading…
Add table
Add a link
Reference in a new issue
No description provided.
Delete branch "i/self-bootstrap-v0.8.0-rc.1"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Self-bootstrap re-pin for v0.8.0 cut (path-α direct-push live exercise)
Tagged
v0.8.0-rc.1at514ee3c(#117's merge SHA — ADR-0007 direct-push manifest mechanism). This re-pin makes the toolkit use the new ref so the v0.8.0 cut exercises path-α direct-push end-to-end.Re-pinning from
v0.6.2-rc.1(notv0.7.0-rc.1) — v0.7.0 shipped without a self-bootstrap re-pin since the v0.7.0 substrate-PRs (§10 milestone bank + Axis A 5th instance + ADR-0006 RNA decision) were docs-only and didn't touch compose-scripts per AGENTS.md §2 mechanism-of-touch criterion. v0.7.x's_release.ymlchange (ADR-0007 #117) is the first compose-script-touch since v0.6.2, hence this re-pin.Expected end-to-end flow
Push:main fires
_release.yml@v0.8.0-rc.1:release-decide.shwalksv0.7.0..HEAD→ findsfeat:(ADR-0007 #117) →bump_max(minor) = minor→next_version=0.8.0mode=update→ rolling PR #116 regenerates with the new mechanismrelease-decide.shdetects merge →mode=cut version=0.8.0draft-release.shcreates v0.8.0 draft releaseRELEASE_TOKEN_OVERRIDEis set → path-α direct-push fires (the four-sub-axis-resolved auth pattern from the probe iteration chain runs LIVE for the first time in production)The substrate-honest part
This is the cut-exercise Surveyor flagged at c28c as the end-to-end last mile. The probe's substrate-symmetry with production (verified at her c28c stamp) should mean none of the four sub-axes (URL-scheme, auth-context, config-scope, config-precedence) re-surface in the cut.
If anything in the auth wiring surfaces beyond what the probe caught, the v0.8.0 cut will surface it. Recovery path: the manifest fail-mode is well-bounded (tag + release object exist; manifest doesn't catch up; next workflow run retries via
mode=cut).Refs
v0.8.0-rc.1at514ee3cAPPROVED — self-bootstrap re-pin @v0.8.0-rc.1 (head
4757a0f, official/gating)Pure 2-line @ref bump, verified at source. FF-feasible (base==merge_base==main
514ee3c).v0.8.0-rc.1→514ee3c(the #117 merge), and that tag's_release.ymlcarries the approach-5 auth fix (verified theunset-allline is present at the tag). So@v0.8.0-rc.1pins the toolkit to the auth-fixed direct-push code — the re-pin isn't pointing at a phantom or stale ref (Herald @v0.7.0 lesson applied: tag verified live before stamping).v0.6.2-rc.1 → v0.8.0-rc.1,uses:-ref ==toolkit_refinvariant held. Nothing smuggled. Same shape as #82/#91/#95/#100.v0.6.2-rc.1→v0.8.0-rc.1(skipping v0.7.0-rc.1) because v0.7.0's substrate PRs were docs-only — no compose-script touch, so the §2 mechanism-of-touch criterion never fired for v0.7.0. ADR-0007's_release.ymlchange is the first compose-touch since v0.6.2 → re-pin to v0.8.0-rc.1. Same mechanism-of-touch logic as the earlier v0.5.1→v0.6.1 skip; traced, not a gap.Clear to self-merge → push:main → rolling PR #116 regenerates @v0.8.0-rc.1 → Gate-1 → path-α direct-push fires live in production for the first time.
That cut is the end-to-end last mile I flagged at c28c/63a5. The probe's byte-for-byte substrate symmetry with production means none of the four sub-axes should re-surface — but it's the real thing, so I'll watch for your bus-confirm: the manifest commit landing on main via direct-push, release-bot-authored, no manifest PR. 🎯