fix(kyverno): let the reports hotfix lint gate move to Go 1.27 #5492

Activity

1. github-project-automation

   added this to 🌊 Project Boardon Oct 28, 2026
2. devantler

   added the

   Bug

   issue type on Oct 28, 2026
3. devantler

   moved this to 📥 Backlog in 🌊 Project Boardon Oct 28, 2026
4. devantler commented on Oct 28, 2026

   

   devantler

   on Oct 28, 2026

   ContributorAuthor

   More actions

   🤖 Generated by the Agentic Engineer

   Decision (ours to make): option 1. The gate exists to catch defects in our watch-plan patch, not to restyle upstream Kyverno. Lint only the packages the patch changes, and keep the full-source go test step as it is. Next: implement this on platform#5483 and show the gate still fails on a finding planted in a patched file.
5. coderabbitai commented on Oct 28, 2026

   

   coderabbitaiBot

   on Oct 28, 2026 – with coderabbitai

   Contributor

   More actions

   ⚠️ Possible Duplicate Issue(s)

   • Reports-controller lint fails with the Go 1.27 build image #5484

   ---

   📝 Issue Planner

   _{Check the box below or use the @coderabbitai plan command to generate an implementation plan and prompts that you can use with your favorite coding assistant.}

   • Create Plan

   ---

   🧪 Issue enrichment is currently in open beta.

   You can configure auto-planning by selecting labels in the issue_enrichment configuration.

   To disable automatic issue enrichment, add the following to your .coderabbit.yaml:

   issue_enrichment:
     auto_enrich:
       enabled: false

   💬 Have feedback or questions? Drop into our discord!
6. devantler commented on Nov 1, 2026

   

   devantler

   on Nov 1, 2026

   ContributorAuthor

   More actions

   🤖 Generated by the Agentic Engineer

   Implementing the 10-28 decision, with one change to its scope. Linting only the patched package is not enough: 2 of the 12 findings are in pkg/controllers/report/resource/controller.go, on upstream lines the patch does not touch. So platform#5483 now limits the gate to the lines the patch changes, within that package.

   Checked locally, on the pinned Kyverno source with the patch applied and golangci-lint v2.15.0, built with Go 1.27.3 rather than the 1.27.2 builder:

   • The tree matches the pinned patched source hash, and the gate reports 0 findings. The same package without the patch filter reports those 2 findings.
   • A wrong fmt.Errorf argument planted in watch_kinds.go fails the gate. So does a formatting fault on a line the patch adds there, and one on a line it changes in controller.go.

   Limit: a line outside the patch is not linted, including a finding that a patched line causes elsewhere. The full go test step is unchanged, and go vet only runs here through the linter. CI on the PR is the check on the real builder.
7. devantler commented on Nov 2, 2026

   

   devantler

   on Nov 2, 2026 – with ChatGPT Codex Connector

   ContributorAuthor

   More actions

   🤖 Generated by the Agentic Engineer

   Consolidating the earlier reader-compatibility report #5484 here. Both track the same Go 1.27 reports-controller upgrade in #5483. #5484 captured the original v2.11.4 export-data failure; this issue records the later v2.15.0 trial and its remaining findings, so maintaining two parallel upgrade plans would duplicate the same work.

   Retained closure criteria: preserve the original failing native job and before/after evidence; record the relevant upstream release review; complete the real lint, test and build stages on the actual selected pinned builder; preserve failure on genuine findings; then complete current-head CI, substantive review, protected merge, default-branch verification and revalidation of #5483. The existing local-versus-native builder difference and line-filter limitation remain explicit and unresolved by this consolidation. This note approves no change to enforcement or lint scope.

   #5492 remains the open owner for the decision and delivery. #5490 separately retains the required-check and merge-queue enforcement defect.