Repository navigation
✨ review: delta re-review mode for PRs already judged at an earlier head - #14622
Conversation
|
[APPROVALNOTIFIER] This PR is NOT APPROVED This pull-request has been approved by: The full list of commands accepted by this bot can be found here. DetailsNeeds approval from an approver in each of these files:Approvers can indicate their approval by writing |
|
@clubanderson is auto-promoted for contributor-lane merging on this PR: the last 5 merged PRs by this author in this hive's repositories have no revert or sentinel-alert evidence. CI, reviewer verdict, diff fences, soak, holds and branch protection still apply. |
Open PR overlap checkBase branch: Conflicts are checked with every PR merged onto the current tip of
|
|
@clubanderson this PR is held at Failing checks:
Next: the hive does not push to your branch. Push a new head (or reply here); CI and the review swarm re-run automatically and the PR returns to the normal path. |
|
@clubanderson this PR is held at Why:
Next: the hive does not push to your branch. Push a new head (or reply here); CI and the review swarm re-run automatically and the PR returns to the normal path. |
A re-review of a PR the hive already judged at another head SHA now carries the previous verdict and findings and asks the reviewer to read only the compare between the two heads. An empty or trivial delta (under review.delta_trivial_lines, touching no previously flagged file) re-stamps the previous verdict onto the new head without a model call. Full review stays for first reviews, prior reject/incomplete, and stale changes_requested findings older than review.delta_max_age. Aggregates record delta.previous_head and delta.trivial so the cost dashboard can split delta reviews from full ones. The new review.delta_* keys are allowlisted in settings parity until a Settings control exists. Fixes #14621 Signed-off-by: Andrew Anderson <andy@clubanderson.com>
385b355 to
f558cd0
Compare
|
Thank you for your contribution! Your PR has been merged. We'd love to hear how your experience was: share feedback |
|
merged by Hive trusted-author auto-merge: author @clubanderson holds role owner and GitHub permission admin on repo; CI green at f558cd0. |
|
/approve |
|
/lgtm |
Fixes #14621
What
When a PR already has a recorded verdict at an earlier head SHA, the re-review kick is a delta re-review: it carries the previous head, date, aggregate and per-perspective verdicts and the findings by file:line, and instructs the reviewer to read only
gh api repos/O/R/compare/<prev>...<new>(plus re-check each previous finding), with the verdict still covering the whole PR at the new head.When the governor's compare shows the delta is empty (identical tree) or under
review.delta_trivial_linestouching no previously flagged file, no reviewer is kicked: the prior aggregate is re-stamped onto the new head with adeltablock and a reason, recorded inreview-verdicts.json, and logged.Full review stays for: first review, prior
reject/incomplete, priorchanges_requestedwith findings older thanreview.delta_max_age,delta_rereview: false.Config (
docs/review-swarm.md)review.delta_rereview(default on)review.delta_trivial_lines(default 20; negative = identical trees only)review.delta_max_age(default 72h)Evidence
~/hive-work/reviewer-spend/summary.md, Oct 13-27: re-reviews were 53% of Bluefin's attributed reviewer spend ($979 of $1,850) and 28% of hive's ($2,019).Recording
Aggregate.Delta{previous_head, trivial, changed_lines, note};PerspectiveReport.delta_previous_headstamped from dispatch state (same path asreview_model), so the cost dashboard can split delta reviews / re-stamps / full reviews.Changes
pkg/review/delta.go(new): planning rules, prompt section, re-stamp,RecordDeltaRestampspkg/review/{dispatch,prompts,review,model_stamps,review_heads}.go: plumbingpkg/github/review_delta.go(new):ReviewDeltaDiffvia CompareCommits (one 300-file page; larger/failed = unknown, no re-stamp)cmd/hive/main.go:planReviewDispatchWithfetches the compare only for candidates (no verdict, no pending, prior at another head) and records re-stampspkg/config: keys, defaults, validation