chore(deps): update ferrlabs/.github digest to e89ec12 - #218
chore(deps): update ferrlabs/.github digest to e89ec12#218ferrlabs-renovate[bot] wants to merge 1 commit into
Conversation
324045d to
86cb199
Compare
86cb199 to
19294c7
Compare
19294c7 to
244bd7a
Compare
244bd7a to
cf1a950
Compare
cf1a950 to
0537d86
Compare
There was a problem hiding this comment.
Two digest/version bumps, both uneventful for this repo.
FerrLabs/.github 4215a84→ce6349b: pulls in two upstream fixes — (1) reusable-security-scan.yml now defaults run-on-private: true and adds a scanned gate that fails if every scan job was skipped. security-scan.yml:9 already sets run-on-private: true, and this run's "Scans ran" check is green, so no behavior change here. (2) reusable-ci-go.yml's sonarqube job now requests job-level pull-requests: write so it can comment on PRs.
Nit: that second fix needs the caller to also grant the permission — the reusable workflow's own comment says so explicitly ("l'appelant doit accorder la même permission... sinon il n'y a rien à transmettre"). ci.yml's ci job (which sets enable-sonar: true) has no permissions: block, so nothing is passed down and the SonarQube job's PR-comment step will still fail to get write access. Not a regression from this bump though — before this fix the job had no permission override at all and was capped at contents: read by the reusable workflow's own top-level permissions: block, so it was already broken pre-bump. To actually benefit from the upstream fix, add to .github/workflows/ci.yml:
ci:
name: CI
permissions:
contents: read
pull-requests: write
uses: FerrLabs/.github/.github/workflows/reusable-ci-go.yml@ce6349be15283fedc44fc17f74555f3b598ebbab # main
FerrLabs/FerrFlow v6.1.1→v6.2.0: adds workspace.releaseCommitBody (opt-in changelog-in-commit feature). No breaking changes, and publish.yml:61 only uses mode: publish, which isn't affected.
CI: gitleaks/osv-scanner/zizmor/opengrep/snyk/"Scans ran"/Conventional commits are all green on this head commit; kind smoke test and the CI (lint/test/build/SonarQube) job hadn't reported yet at review time — no failure signal from the bump so far.
0537d86 to
4e1c1ee
Compare
SonarQube — aucune nouvelle issueComparaison entre le projet bac à sable de cette PR et la branche par défaut : SonarQube Community n'analyse pas les PR, ce delta est calculé côté CI. Détail |
4e1c1ee to
3199c25
Compare
3199c25 to
7df811b
Compare
7df811b to
72c1286
Compare
There was a problem hiding this comment.
Follow-up on this rebased head. Two things changed since my last review, both good:
- The Nit I raised (missing
pull-requests: writepassthrough for the SonarQube reusable job) is now fixed —ci.ymlonmainpicked up FerrVault#222 (432a7211), which added thepermissions:block. This PR's rebase carries it, and it's not part of this PR's own diff. FerrLabs/.githubdigest moved further (ce6349b→f14308c) as Renovate kept rebasing while the PR sat open. Checked the intervening commits (#219–#226 on FerrLabs/.github): all self-referential digest bumps and renovate-config/Kellnr changes, nothing touchingreusable-ci-go.yml,reusable-ferrflow-release.yml,reusable-renovate-dispatch.yml,reusable-sbom-track.yml, orreusable-security-scan.yml. No new behavior for this repo.
FerrFlow bump (v6.1.1→v6.2.0) is unchanged from before: adds opt-in workspace.releaseCommitBody, not used by publish.yml:61 (mode: publish only).
CI was still pending/in-progress at review time (no failures observed); scan jobs (gitleaks/osv-scanner/zizmor/snyk) are green as before. No blocking findings.
72c1286 to
801d831
Compare
e4b35b0 to
6c78bce
Compare
There was a problem hiding this comment.
Follow-up on this rebased head. Since my last review (f47ab42d, FerrFlow v7.0.5 / .github digest 67e9379), the branch has been rebased repeatedly and both pins moved a lot further:
- FerrFlow direct pin (
publish.yml): v7.0.5→v7.8.1. Checked every release note in between (v7.1.0–v7.8.1) — noBREAKINGentries, all features (latestTag, api-check, pypi publisher, versionTemplate, config include, plan --interactive, etc.) or bug fixes.publish.yml:61still only usesmode: publishwith docker+helm publishers (per.ferrflow); none of the npm/pypi/OIDC-publish changes apply. The one item worth naming: v7.3.1 added npm trusted publishing over OIDC — irrelevant here since this repo has no npm publisher. .githubdigest:67e9379→93b84d0, pulling the same FerrFlow range intoreusable-ferrflow-release.yml(now pinned v7.8.0, was v7.0.2) plus unrelated repo-hygiene commits (LF normalisation, renovate docker-auth fix, cargo--locked/env-flag CI tweaks, github-script v9). Diffedreusable-ci-go.yml(byte-identical) andreusable-security-scan.yml(only thecodeql-action/upload-sarifpin moved, same v4) between the two digests — no behavior change reaches this repo.- The FerrFlow v7
commitFormatsleniency Nit from my earlier review is still open and unchanged — not repeating it.
CI: all 21 checks green on this head commit (Build, Test, Lint, Security, SonarQube, kind smoke test, gitleaks/osv-scanner/zizmor/opengrep/snyk, PR title). No blocking findings.
6c78bce to
c919180
Compare
There was a problem hiding this comment.
Follow-up on this rebased head (f47ab42d→c919180). Changes since my last review:
-
FerrFlow direct pin (
publish.yml): v7.0.5→v7.8.1 — a run of additive minor features (versionTemplate,workspace.latestTag, version-source reporting, per-package configinclude,api-check, a pypi publisher,graph, interactiveplan, repeatable--force-version/--exclude) plus bug fixes. No breaking changes.publish.yml:61still only usesmode: publish, unaffected by any of them. -
FerrLabs/.githubdigest moved67e9379→d7083e2, which bumps the FerrFlow pin insidereusable-ferrflow-release.yml(backs this repo'srelease.yml) in lockstep, v7.0.2→v7.8.1 — now aligned with the direct pin. Checked that workflow: FerrVault's release job doesn't setmode: pr, so v7.8.0's "tag on merge instead of the pre-bump commit" PR-mode fix is a no-op here.
The FerrFlow v7 commitFormats leniency Nit from my first review is still open and unchanged — not repeating it.
CI: all 19 checks green (Build, Test, Lint, Security, SonarQube, kind smoke test, gitleaks/osv-scanner/zizmor/opengrep/snyk, PR title). No blocking findings.
de79606 to
7e411ba
Compare
7e411ba to
24d893a
Compare
24d893a to
a02c52b
Compare
There was a problem hiding this comment.
Follow-up on this rebased head. Since my last review (commit f47ab42d), the FerrLabs/.github digest moved 67e9379→c34090d (~30 upstream commits, Aug 15–27).
Diffed all 6 reusable workflows FerrVault calls at both SHAs:
reusable-ci-go.yml,reusable-pr-title.yml,reusable-renovate-dispatch.yml,reusable-sbom-track.yml— byte-identical, no changes.reusable-security-scan.yml— only a patch-levelcodeql-action/upload-sarifdigest refresh across the 5 SARIF-upload steps. No input/permission/runner changes.reusable-ferrflow-release.yml(backsrelease.yml) — the internalFerrLabs/FerrFlowaction pin moved v7.0.2→v7.9.0. Checked all 20 intervening release notes: noBREAKING CHANGEentries. The few behavior-relevant changes (newworkspace.versionSourceconfig defaulting to today'shighestbehavior; extended token-stripping for non-GitHub hosts; areleaseCommitMode: "pr"tag-timing fix) all require opt-in config or non-GitHub hosting that FerrVault doesn't use — confirmed FerrVault has noferrflow.json/.ferrflow.json, so it runs on defaults throughout.
Net: this is a routine transitive digest refresh, nothing new for FerrVault's mode: release usage beyond the already-flagged (and still-open, not repeating) v7.0.0 commitFormats leniency Nit.
CI: gitleaks/osv-scanner/zizmor/snyk/opengrep/Scans-ran/Lint/Test/Security/PR-title all green on this head. Build, SonarQube analysis, and kind smoke test were still in progress at review time with no failures observed. No blocking findings.
There was a problem hiding this comment.
Follow-up on this rebased head. FerrLabs/.github digest moved 67e9379→62a681d (~29 upstream commits over 12 days). Checked all 5 reusable workflows FerrVault actually uses:
reusable-ci-go.yml,reusable-renovate-dispatch.yml: byte-identical, no change.reusable-pr-title.yml:runnerinput default changed from hardcodedferrlabs-k8s-lightto auto-select (falls back toubuntu-latestfor public repos instead of hanging). FerrVault doesn't overriderunner, so this only helps (or is a no-op if private).reusable-ferrflow-release.yml: only its internalFerrLabs/FerrFlowaction pin moved v7.0.2→v7.9.0. Checked all 8 intervening release notes — no new BREAKING CHANGE entries beyond the already-known v7.0.0commitFormatsleniency Nit (previously reported, still open, not repeating it).reusable-security-scan.yml: only acodeql-action/upload-sarifdigest refresh within v4, no input/default changes;run-on-privatestill defaultstrue, matchingsecurity-scan.yml:9.
Also checked commit #266 (removes a reusable SBOM-publish workflow) — FerrVault never referenced it (sbom doesn't appear in any of its workflows), so not applicable.
CI: all checks green (Build, Test, Lint, Security, SonarQube, kind smoke test, gitleaks/osv-scanner/zizmor/opengrep/snyk, PR title). No blocking findings.
There was a problem hiding this comment.
Follow-up on this rebased head. Since my last review (commit f47ab42d), FerrLabs/.github digest moved 67e9379→34983fe (30 commits, #260→#300):
- FerrFlow pin inside
reusable-ferrflow-release.yml(backsrelease.yml) jumped v7.0.2→v7.11.1. Checked all 10 intervening minor releases (v7.1.0–v7.11.0): all additive features (new CLI subcommands, opt-in config keys, new publishers) and bug fixes, none marked breaking.publish.ymlstill only usesmode: publish, unaffected. The v7.0.0commitFormatsleniency Nit I raised earlier is unchanged — not repeating it. reusable-pr-title.ymlgained runner auto-selection by repo visibility — FerrVault'spr-title.ymlalready passesrunner: ubuntu-latestexplicitly, so this is a no-op here.- The removed reusable SBOM-publish workflow and the Rust cargo-flag/
--lockedchanges target files/workflows FerrVault doesn't use (nosbom.yml, andci.ymlusesreusable-ci-go.yml, not the Rust one).
No blocking findings. CI: all relevant checks green (Lint, Security, gitleaks/osv-scanner/zizmor/snyk/opengrep, PR title); Test and kind smoke test still in progress at review time with no failures observed.
There was a problem hiding this comment.
Follow-up on this rebased head. Since my last review (commit f47ab42), FerrLabs/.github digest moved 67e9379→340c925, pulling in ~30 commits. Diffed the five reusable workflows this repo actually calls between those two SHAs:
- reusable-ci-go.yml: only the
detectjob's default private-repo runner moved toferrlabs-k8s-light(light runner, no build). No effect onlint/test/build. - reusable-security-scan.yml: only the
upload-sarifaction digest bumped (codeql-action v4). No behavior change. - reusable-pr-title.yml: the
runnerinput's default changed (ferrlabs-k8s-light→ auto-select by visibility), but FerrVault'spr-title.ymlalready pinsrunner: ubuntu-latestexplicitly, so this repo is unaffected either way. - reusable-ferrflow-release.yml: FerrFlow pin bumped v7.0.2→v7.11.1 (backs
release.yml). Read all 30 intervening release notes (v7.0.3 through v7.11.2) — all features are opt-in/config-gated (versionTemplate, api-check, pypi publisher, rollback, calver, etc.) or internal fixes (git/publish/docker robustness). No default-behavior break found; the v7.0.0commitFormatsleniency Nit I raised earlier is unrelated and unchanged. - reusable-renovate-dispatch.yml: byte-identical, no change.
No blocking findings. CI: all completed checks green (Lint, Test, Security, gitleaks/osv-scanner/zizmor/snyk/opengrep, Scans ran, PR title); kind smoke test and CI / Build + CI / SonarQube were still in progress at review time with no failures observed.
There was a problem hiding this comment.
Follow-up on this rebased head. Since my last review (commit f47ab42), FerrLabs/.github digest moved 67e9379→a5886d74. Checked every intervening commit (#282–#303) for anything touching the reusable workflows this repo calls:
reusable-ferrflow-release.yml: the pinned FerrFlow action moved v7.0.2→v7.11.1. Read every release note in between (v7.0.3 through v7.11.1) — all additive opt-in features (workspace.latestTag,versionTemplate,ferrflow graph,ferrflow rollback, per-package config, etc.) or bug fixes; no default-behavior or breaking change.release.ymlonly passesdry-run, unaffected. The previously-flagged Nit (v7.0.0's more lenientcommitFormatsdefault) is unchanged in this range — not repeating it.reusable-pr-title.yml(#295/#293): now auto-selects the runner by repo visibility instead of alwaysubuntu-latest.pr-title.yml:23already pinsrunner: ubuntu-latestexplicitly, so this repo is unaffected.reusable-security-scan.yml(#303): opengrep ruleset is now fetched via HTTPS/tarball instead ofgit clone— internal implementation detail, no input/output change. Confirmed green (Secrets + CVE / opengrep (SAST)).- #302: five no-build helper jobs (across
reusable-ci-go.yml,reusable-ci-rust.yml,reusable-docker-build.yml,reusable-ferrflow-release.yml) moved to the light runner tier — pure infra/cost change.
Nothing else in the range (#282–#303) touches a workflow FerrVault calls.
CI: all 19 checks green (Build, Lint, Test, Security, SonarQube, kind smoke test, gitleaks/osv-scanner/zizmor/opengrep/snyk, Scans ran). No blocking findings.
There was a problem hiding this comment.
Follow-up on this rebased head. Since my last review (FerrLabs/.github digest 67e9379), the digest moved further to 71acd9a (~30 commits). Checked every commit in that range against the 5 reusable workflows FerrVault actually uses:
reusable-ci-go.yml,reusable-renovate-dispatch.yml: no caller-visible change (reusable-renovate-dispatch.ymluntouched entirely; ci-go's only change is a private-repo runner-tier switch for thedetectjob, same job name).reusable-pr-title.yml:runnerinput default changed to auto-select, but FerrVault'spr-title.ymlpassesrunner: ubuntu-latestexplicitly, so the new fallback logic never triggers — no effect.reusable-ferrflow-release.yml: pinned FerrFlow action moved v7.0.2→v7.11.1. Checked all 25 intervening release notes — Features/Bug Fixes only, no new BREAKING entries beyond the already-flagged v7.0.0commitFormatsleniency (still open, not repeating it).reusable-security-scan.yml: opengrep ruleset fetch switched fromgit clonetocurl codeload.github.com+tar— same output, no new input/secret.
No renamed/removed jobs, no new required inputs/secrets/permissions, no required-check renames — nothing here should break FerrVault's CI or branch protection.
Nit: FerrFlow v7.9.0 switched release-commit authoring to GraphQL createCommitOnBranch (verified commits) — changelog doesn't call it breaking, but worth a quick sanity check that FerrVault's next release run still produces the expected signed commit.
Nit: the opengrep fetch now needs codeload.github.com reachable instead of github.com — if FerrVault's self-hosted runner has an egress allowlist, worth confirming it covers the new host.
CI: gitleaks/osv-scanner/zizmor/snyk/opengrep/Scans ran/Lint/Test/Security/kind smoke test all green on this head commit. Build and SonarQube analysis were still in progress at review time, no failures observed. No blocking findings.
There was a problem hiding this comment.
Follow-up on this rebased head — a large gap since my last review (f47ab42d, 2026-08-16), so re-checked the full delta.
FerrLabs/.github digest 67e9379→3df97ce (17 days, ~25 commits): the only functionally relevant changes for FerrVault's callers:
reusable-ci-go.yml'ssonarqubejob now delegates to a separatereusable-sonarqube-scan.yml@2883414f(was inline before), still requiring job-levelpull-requests: writepassthrough from the caller.ci.yml:9-12already carries that block (added back in FerrVault#222, confirmed present at this head), and this run's "CI / SonarQube" check is green — no action needed.reusable-ferrflow-release.yml's FerrFlow pin (backsrelease.yml) advanced v7.0.2→v7.11.1. Checked every release note v7.0.5→v7.15.1 forBreaking Changes: none — the only breaking entry in range is the v7.0.0commitFormatsleniency I already flagged as a Nit previously (unchanged, not repeating).reusable-pr-title.yml,reusable-security-scan.yml(opengrep ruleset now fetched via HTTPS tarball instead of git clone, some helper jobs moved to a lighter runner tier),reusable-renovate-dispatch.yml: no changes affecting this repo's usage.
No blocking findings. CI: all checks green (Build, Test, Lint, Security, SonarQube, kind smoke test, gitleaks/osv-scanner/zizmor/opengrep/snyk, PR title).
| pull-requests: write | ||
| name: CI | ||
| uses: FerrLabs/.github/.github/workflows/reusable-ci-go.yml@4215a840974a9eb611b05f1496ec99d7ba0492c0 # main | ||
| uses: FerrLabs/.github/.github/workflows/reusable-ci-go.yml@e89ec1209ae561a7a1b910127ccbd926e77fa613 # main |
| jobs: | ||
| dispatch: | ||
| uses: FerrLabs/.github/.github/workflows/reusable-renovate-dispatch.yml@a4b4a70d116127cd49f09ad6f363ac55fd8f6501 # main | ||
| uses: FerrLabs/.github/.github/workflows/reusable-renovate-dispatch.yml@e89ec1209ae561a7a1b910127ccbd926e77fa613 # main |
| scan: | ||
| name: Secrets + CVE | ||
| uses: FerrLabs/.github/.github/workflows/reusable-security-scan.yml@4215a840974a9eb611b05f1496ec99d7ba0492c0 # main | ||
| uses: FerrLabs/.github/.github/workflows/reusable-security-scan.yml@e89ec1209ae561a7a1b910127ccbd926e77fa613 # main |
There was a problem hiding this comment.
Follow-up on this rebased head. Since my last review (commit f47ab42d, digest 67e9379), FerrLabs/.github digest moved to e89ec1209ae561a7a1b910127ccbd926e77fa613. The notable change: the FerrFlow action pinned inside reusable-ferrflow-release.yml (backs this repo's release.yml) jumped v7.0.2→v7.11.1. Read all intervening release notes (v7.1.0–v7.11.1) — all minor/patch features and bug fixes, no breaking changes listed. The two entries that touch behavior gated behind flags this repo doesn't set (fix(release): pr mode tag on merge #936, needs mode: pr; the trigger-renovate job's runner relabel) don't apply — FerrVault's release.yml uses default release mode + mode: publish, no deferred-publish/trigger-renovate.
Other changes in this digest bump, both no-op for FerrVault:
reusable-security-scan.yml: opengrep ruleset fetch switched fromgit clonetocurl+tar(upstream reliability fix), plus a routinecodeql-action/upload-sarifdigest bump.reusable-ci-go.yml/ the release workflow'strigger-renovatejob: private-repo runner relabeledferrlabs-k8s→ferrlabs-k8s-light(infra-side, not this repo's concern).
The FerrFlow v7 commitFormats leniency Nit I raised earlier is still open and unchanged — not repeating it here.
CI: all green (Lint, Test, Security, gitleaks/osv-scanner/zizmor/opengrep/snyk, Scans ran, Conventional commits) except CI / Build, CI / SonarQube analysis, and kind smoke test, still in progress at review time with no failures observed. No blocking findings.
This PR contains the following updates:
4215a84→e89ec12Configuration
📅 Schedule: (UTC)
🚦 Automerge: Enabled.
♻ Rebasing: Whenever PR is behind base branch, or you tick the rebase/retry checkbox.
👻 Immortal: This PR will be recreated if closed unmerged. Get config help if that's undesired.
This PR has been generated by Mend Renovate CLI.