Repository navigation
ci: repin Packit SRPM image to live digest - #85
Conversation
Update quay.io/packit/packit digest in packit-srpm-pilot.yml from the pruned sha256:149e6e06... digest to the active manifest-list digest sha256:8a1784251c51eed7a094820c894e2ee7f4ed4bbce4eb78eb172a04de3fae43e1. Also update documentation references, add unit test assertions, and document exit 125 manifest unknown triage guidance in the build-failure-triage skill. Fixes projectbluefin#84 Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Signed-off-by: Danathar <Danathar@users.noreply.github.com>
hanthor
left a comment
There was a problem hiding this comment.
Verified independently against the live registry and against a real CI log. The change is correct.
The digest on main is genuinely gone, and the replacement is genuinely live. Queried quay.io directly with a pull token:
--- sha256:8a1784251c51eed7a094820c894e2ee7f4ed4bbce4eb78eb172a04de3fae43e1 ---
HTTP 200
--- sha256:149e6e06d3e5fb2f10d19760c8a0031c7d8825e7bb91a5f4a7ab9b927c947494 ---
HTTP 404
And that 404 is what is actually breaking CI today. From the srpm (zsh) job on PR #72's run (job 103321271752):
Unable to find image 'quay.io/packit/packit@sha256:149e6e06d3e5fb2f10d19760c8a0031c7d8825e7bb91a5f4a7ab9b927c947494' locally
docker: Error response from daemon: manifest for quay.io/packit/packit@sha256:149e6e06... not found: manifest unknown: manifest unknown
##[error]Process completed with exit code 125.
That is the exact exit-125 / manifest unknown signature your build-failure-triage SKILL.md update describes, so the skill change is grounded too.
Gates, run locally on this branch merged into main (54cc8d1):
PR#85 fc=0 val=0 pytest=0 | validated 193 source RPMs | 72 passed in 0.62s
PR#85 quoting=0 runtime_contract=0 unittest=0 | validated runtime contract: 72 packages
Identical to the unmodified-main baseline, so nothing regressed.
This supersedes #82. #82 proposes the identical one-line digest swap and nothing else. This PR additionally pins the digest with assertions in tests/test_packit_srpm.py (so the next prune fails a cheap unit test instead of the SRPM matrix) and updates the two docs that quote the old digest. Keep this one, close #82.
One thought, not a blocker: pinning a digest that upstream prunes on a schedule will recur. A follow-up that resolves quay.io/packit/packit:stable at job time and records the digest in the step summary, or a Renovate rule for this image, would stop the next occurrence. (Note :stable currently 404s too — :latest or whatever tag Packit actually publishes would need checking first.)
Generated by Claude Code
Summary
quay.io/packit/packitin.github/workflows/packit-srpm-pilot.ymlfrom pruned digestsha256:149e6e06d3e5fb2f10d19760c8a0031c7d8825e7bb91a5f4a7ab9b927c947494to the active manifest-list digestsha256:8a1784251c51eed7a094820c894e2ee7f4ed4bbce4eb78eb172a04de3fae43e1.docs/architecture.mdanddocs/superpowers/specs/2026-09-05-full-packit-rpm-factory-design.md.tests/test_packit_srpm.pyverifying the live digest is pinned and the pruned digest is not present..agents/skills/build-failure-triage/SKILL.mdto classify exit 125manifest unknownas upstream digest pruning requiring repinning.Fixes #84
Validation
python3 -m unittest discover -s tests(72 passed)python3 tools/check_workflow_quoting.pypython3 tools/validate.pypython3 tools/factory_contract.pysha256:8a1784251c51eed7a094820c894e2ee7f4ed4bbce4eb78eb172a04de3fae43e1via quay.io API.— hive: backend=agy
🐝 Hive Agent:
contributor| SHA:15a9dd5