Skip to content

feat(review): high-signal reviews + visible Gate re-run on checkbox - #1091

Merged
JSONbored merged 1 commit into
mainfrom
feat/review-focus-and-rerun-visibility
Jun 23, 2026
Merged

feat(review): high-signal reviews + visible Gate re-run on checkbox#1091
JSONbored merged 1 commit into
mainfrom
feat/review-focus-and-rerun-visibility

Conversation

@JSONbored

Copy link
Copy Markdown
Owner

Follow-up to the blockers/nits depth port (#1089), which is now live and producing real finding-by-finding reviews (e.g. metagraphed #1528 correctly closed on a genuine missing-import ReferenceError).

1. Noise → high signal

The deep review over-listed: ~9 blockers (many the same 'doesn't handle empty array / catch errors' pattern repeated per function), ~12 nits, and a 'why blocked' that repeated every blocker. Tuned:

  • Prompt: be SELECTIVE (≤~3 blockers / ~5 nits), deduplicate recurring issues, tighten severity (defensive 'should handle X' hardening is a nit, not a blocker, unless a real input will actually fail), concise 2–4 sentence assessment.
  • Caps: parse 12→6; rendered blockers 8→3, nits 12→5.
  • Gate reason: the consensus defect's 'why blocked' cites only the primary blocker, not the whole list joined.

2. Re-run visibility (task B)

Clicking 'Re-run Gittensory review' now flips the Gittensory Gate check to in_progress before the re-review runs, so it visibly goes running → concluded like a real CI step (fail-safe; resolves the same run by name+headSha, no duplicate).

Full suite green (3485). Internal to the review path; gate/comment contracts unchanged.

Two review-quality fixes after the blockers/nits depth port went live:

1. Noise: the deep review over-listed findings (e.g. 9 blockers, many the same
   'doesn't handle empty array / catch errors' pattern repeated per function, plus
   a 'why blocked' that repeated every blocker). Tune for high signal:
   - prompt: be SELECTIVE (~3 blockers / ~5 nits max), DEDUPLICATE recurring issues,
     and tighten severity (defensive 'should handle X' hardening is a NIT, not a
     blocker, unless a real input will actually fail).
   - caps: parse 12->6; rendered blockers 8->3, nits 12->5.
   - consensus defect: the Gate's 'why blocked' cites only the PRIMARY blocker, not
     the whole list joined together.

2. Re-run visibility (B): clicking 'Re-run Gittensory review' now flips the Gittensory
   Gate check to in_progress before the re-review runs, so it visibly goes running ->
   concluded like a real CI step (fail-safe; resolves the same run by name+headSha).
@dosubot dosubot Bot added the size:M label Jun 23, 2026
@codecov

codecov Bot commented Jun 23, 2026

Copy link
Copy Markdown

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 94.76%. Comparing base (2c35207) to head (d564b1c).
✅ All tests successful. No failed tests found.

Additional details and impacted files
@@           Coverage Diff           @@
##             main    #1091   +/-   ##
=======================================
  Coverage   94.76%   94.76%           
=======================================
  Files         153      153           
  Lines       18540    18542    +2     
  Branches     6705     6705           
=======================================
+ Hits        17570    17572    +2     
  Misses        418      418           
  Partials      552      552           

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:
  • 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

@JSONbored
JSONbored merged commit 6ff7271 into main Jun 23, 2026
16 checks passed
@JSONbored
JSONbored deleted the feat/review-focus-and-rerun-visibility branch June 23, 2026 18:20
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant