Repository navigation
feat(roadmap-planner): W6 — flip gitlab.hydrate_diff default to true - #192
Conversation
Audit follow-up B12 / Q12 (2026-05-19). Merged-MR diff stats (additions / deletions / changed_files) feed the Dashboard tab and the GitHub side collects them on every cycle. Asymmetry is now fixed: the GitLab default flips on too. - gitlab.Syncer.HydrateDiff initialised to true in NewSyncer. - viper default for `gitlab.hydrate_diff` flipped to true. - TestSyncerHydrateDiffDefault pins the new default so a future refactor doesn't silently flip it back. Cost is ~+1 extra API call per merged MR; ~+2% of the per-cycle GitLab API budget. Operators on tight rate limits can disable via `gitlab.hydrate_diff: false` in the ConfigMap. The draft-skip mirror (W6 design item 2) is already in main — gitlab/sync.go already skips notes for `mr.Draft || mr.WorkInProg`. `changes_requested` review-state marker for GitLab (Q5) is deferred. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
| @@ -0,0 +1,215 @@ | |||
| # Team Analytics — Behaviour Changes | |||
There was a problem hiding this comment.
Suggestion (docs/missing): Consider adding a brief file-level comment indicating this is part of the audit follow-up from 2026-05-19 (e.g., `# Team Analytics — Behaviour Changes
Audited: 2026-05-19
Plan: docs/team-analytics/audit-2026-05-19/PLAN.md`). This helps future readers understand the context.
🤖 AI Code Review
SummaryThis PR flips the default value of Review Statistics
Warnings
Suggestions
Positive Feedback
Review: This PR is ready to merge. The warning is non-blocking but should be addressed for documentation consistency. ℹ️ About this reviewThis review was automatically generated using the
|
Audit follow-up B12 / Q12 — 2026-05-19.
Summary
gitlab.NewSyncerinitialisesHydrateDiff: true(wasfalse).gitlab.hydrate_diffflipped totrue.TestSyncerHydrateDiffDefaultpins the new default.Why
Merged-MR additions / deletions / changed_files feed the Dashboard tab. GitHub already collects them every cycle (returned inline on the list endpoint); GitLab needed a separate show-MR call. The asymmetry meant Dashboard counts only included GitHub-side activity. With
hydrate_diffon by default, both sides feed the panel.Cost is ~+1 extra API call per merged MR; ~+2% of the per-cycle GitLab API budget. Per maintainer (Q12), the one-shot backfill cost is fine to absorb overnight.
Already on main
The "skip review fetch for draft MRs" mirror (W6 design item 2) is already in
gitlab/sync.go— drafts and WIP MRs are skipped viaskipNotes := mr.Draft || mr.WorkInProg. This PR only flips the diff default.Deferred
changes_requestedreview-state marker for GitLab (Q5) — Q5 answer was "defer, we generally don't do that in gitlab."Test plan
go test ./...+go vet ./...— greenTestSyncerHydrateDiffDefaultpins the constructor defaultpull_requests.additions/deletions/changed_filespopulate on the next GitLab cycleRollback
gitlab.hydrate_diff: falsein the ConfigMap.🤖 Generated with Claude Code