Repository navigation
chore(deps): update projectbluefin/actions digest to 8895d09, and freedesktop-sdk to 26.08.1 + gnome-build-meta to 50.4-5 - #183
mergeraptor[bot] wants to merge 3 commits into
Conversation
Edited/Blocked NotificationRenovate will not automatically rebase this PR, because it does not recognize the last commit author and assumes somebody else may have edited the PR. You can manually request rebase by checking the rebase/retry box above. |
There was a problem hiding this comment.
HUMAN DECISION NEEDED — hold or close: this PR fails validation by construction until #180 lands, but Renovate automerge is enabled.
Correctness findings:
-
Merging turns
mainred (high).elements/freedesktop-sdk.bst:8bumps the FSDK junction tofreedesktop-sdk-26.08.1-0-gb02b59f, butproject.conf:29still pinsrelease-version: "26.08.0"..github/scripts/check-release-version.pyfails closed on exactly this drift (it exists to prevent sysupdate assets shipping stale version strings), sojust validate/ "Validate Element Graph" fails. Verified the pin and script at headcabe008. -
Not agent-fixable here. Reverting the two junction hunks won't stick: the
track-refsjob inbuild.ymlre-resolves and pushes refs on everypull_requestpush torenovate/*branches (as documented in the maintainer note; same pattern killed #172/#176). The durable fix is PR #180, which is still open. -
Automerge risk. The Renovate config on this PR has automerge enabled; if required checks ever pass spuriously (or the check is not marked required), a known-red change lands. Suggest a human either close this (Renovate will not re-open per its ignore semantics — confirm that's acceptable) or leave it parked with automerge disabled until #180 merges.
The actions digest bump itself (.github/workflows/build.yml:179, 6680f74 → 8895d09) is fine and can be re-run cleanly after #180.
— hive: agent=reviewer backend=copilot model=claude-fable-5 copilot=1.0.78
|
@castrojo — same situation as #176: read access only here, and the |
|
Blocked on #180, and structurally red until it lands.
#180 fixes the root cause: it moves junction tracking to a scheduled Also note #84, which fixes the Leaving open, not closing: the runner and action-digest hunks here are independently wanted. Please rebase after #180. |
There was a problem hiding this comment.
HUMAN DECISION NEEDED — #180 has merged since this PR was noted as blocked; a maintainer must rebase/re-run (or close and let Renovate reopen) so the contaminated junction hunks drop out.
intent-alignment
- The advertised change is only the
projectbluefin/actionsdigest bump (.github/workflows/build.yml:179), but the diff also carries two junction ref bumps Renovate did not author:elements/freedesktop-sdk.bst:8(26.08.0 → 26.08.1) andelements/gnome-build-meta.bst:8(50.4-0 → 50.4-5), injected by thetrack-refsjob (.github/workflows/build.yml:26-56at head cabe008). - Merging as-is breaks
main:project.conf:29pinsrelease-version: "26.08.0"and.github/scripts/check-release-version.pyfails closed on the 26.08.1 FSDK ref, sojust validategoes red — exactly as the maintainer note states. - The stated blocker, PR #180 ("move junction ref tracking out of the pull_request path"), is now MERGED. This PR's head predates it and still contains the old
track-refsjob and the contaminated hunks; a rebase onto current main should reduce it to the clean digest bump.
No findings from: security, style, docs-currency.
— hive: agent=reviewer backend=copilot model=claude-fable-5 copilot=1.0.78
cabe008 to
e8c913f
Compare
|
The digest and version bumps look right, and the failure is a one-line gap in the same change rather than anything upstream.
That is what both This is the same gap as #176, which is the other PR carrying the FSDK 26.08.1 bump. It may be worth a Renovate post-upgrade task so the pin and |
There was a problem hiding this comment.
correctness (high)
The freedesktop-sdk bump to 26.08.1 (elements/freedesktop-sdk.bst:8) drifts against release-version: "26.08.0" in project.conf:29. .github/scripts/check-release-version.py:69-82 fails closed on exactly this, and it is the first thing just validate runs (build.yml:41-42) — CI on head e8c913f is already red (build: failure, unit: failure). If forced through, the drift the script documents (release assets keep the old version string, so systemd-sysupdate sees no update) would stall the fleet.
intent-alignment (medium)
The Renovate body advertises only the actions digest bump; the two junction-ref hunks (elements/freedesktop-sdk.bst:8, elements/gnome-build-meta.bst:8) are contamination from the old track-refs job, as the maintainer note discloses. One update to that note: it says the hunks "cannot be durably reverted here" pending #180, but #180's fix is already on this branch's head — .github/workflows/track-junctions.yml exists (scheduled-only, no pull_request trigger, lines 14-18) and build.yml no longer contains a track-refs job. Ticking the rebase checkbox should now shed both hunks and leave the clean digest bump, with no re-contamination on push.
The actions digest itself checks out: 8895d09 is a real commit on projectbluefin/actions (chore(deps): update github-actions (non-major) (#508)), full-SHA pinned, and the calling job keeps read-only permissions (build.yml:138-140).
No findings from: security, style, docs-currency.
Confidence: 2/5 (needs attention) — 1 high finding; 1 medium finding
— hive: agent=reviewer backend=copilot model=claude-fable-5 copilot=1.0.78
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
e8c913f to
854e882
Compare
There was a problem hiding this comment.
HUMAN DECISION NEEDED — decide whether an OS point release (FSDK 26.08.1 + gnome-build-meta 50.4-5 + release-version bump) should land inside this automerge-labeled Renovate digest PR, or be re-run as clean PRs per the plan stated in the PR body.
intent-alignment
The PR body and the head contradict each other. The maintainer note says "It cannot merge as-is", "Blocked on #180", and "No code was changed by me. Title and this note only" — but head commit 854e882 (authored by castrojo, after that note) bumps project.conf:29 release-version to "26.08.1" and edits docs/skills/flatcar-parity-matrix.md:58, which unblocks exactly the check the note said blocks the PR (.github/scripts/check-release-version.py passes at head; verified locally). #180 has since merged, and its stated design was that junction tracking gets its own scheduled PR — yet this PR keeps the bot-injected junction hunks (elements/freedesktop-sdk.bst:8, elements/gnome-build-meta.bst:8) and legitimizes them. Net effect: a chore/deps PR with the automerge label now ships a fleet-visible OS release (new 26.08.1 sysupdate assets, per the invariant documented in check-release-version.py). That may be intended, but the body says the opposite, so a human should confirm and update the body before this merges.
docs-currency
docs/skills/flatcar-parity-matrix.md:58now cites junctionfreedesktop-sdk-26.08.1, but the table's "FSDK 26.08 version" column comes from hardcoded pins inscripts/generate-parity-matrix.py(IN_SCOPE_COMPONENTS, e.g. glibc2.44-23at line 55, systemd261.2at line 72) that were audited against 26.08.0. Regeneration only re-stamps the footer; the component versions were not re-verified against 26.08.1, so the doc now implies a re-audit that did not happen. (low)
Verified: FSDK ref 26.08.1-0-gb02b59f… matches upstream tag freedesktop-sdk-26.08.1^{}; gnome-build-meta d78f498… is a real commit on upstream gnome-50; actions digest 8895d09 is on projectbluefin/actions main, and the consumed server-installer-test.yml change between digests is only a pinned taiki-e/install-action bump.
No findings from: correctness, security, style.
Confidence: 3/5 (needs attention) — 1 high finding
— hive: agent=reviewer backend=copilot model=claude-fable-5 copilot=1.0.78
|
Closing: contaminated by legacy pull_request junction tracking with freedesktop-sdk 26.08.1 and gnome-build-meta 50.4-5 hunks, which conflicted with main. Junction tracking was separated onto track-junctions.yml (#180). Renovate will generate a clean digest bump for projectbluefin/actions without junction contamination. |
This PR contains the following updates:
6680f74→8895d09Configuration
📅 Schedule: (UTC)
🚦 Automerge: Enabled.
♻ Rebasing: Whenever PR becomes conflicted, or you tick the rebase/retry checkbox.
🔕 Ignore: Close this PR and you won't be reminded about this update again.
This PR has been generated by Mend Renovate CLI.
Maintainer note (castrojo, PR triage sweep). Retitled to disclose two hunks this PR silently acquired.
Besides the advertised actions digest bump, the diff also carries:
elements/freedesktop-sdk.bst:26.08.0-0-gdb97cce->26.08.1-0-gb02b59felements/gnome-build-meta.bst:50.4-0-g8524013->50.4-5-gd78f498Renovate did not author those.
build.yml'strack-refsjob fires on everypull_requestevent forrenovate/*branches and pushes re-resolved junction refs onto the branch, so they reappear on any push and cannot be durably reverted here.This PR is a live demonstration of the problem: it was opened during a triage sweep in which I had just closed #172 for carrying exactly these two hunks, and it was contaminated immediately.
It cannot merge as-is.
project.confpinsrelease-version: "26.08.0"and.github/scripts/check-release-version.pyis the first thingjust validateruns, failing closed on drift. Landing FSDK 26.08.1 against a declared 26.08.0 turnsmain's build red — the same two-to-four-second "Validate Element Graph" failure seen on #172 and #176.Blocked on #180, which moves junction tracking to a scheduled workflow with its own PR and syncs
release-versionto the tracked FSDK point release in the same commit. Once that lands, this becomes a clean digest bump and can be re-run.No code was changed by me. Title and this note only.