Repository navigation
fix(dashboard): standup and pr-watch count the items the work fold forgot - #22856
Conversation
UX Review (Opus 5.5) — ⏭️ skippedRevision |
First Principles Review (Opus 5.5) — ✅ PASSPremise-level review of First-Principles-Verdict: PASS The version bump contradicts the description, and the new spec paragraph repeats the one above it. Neither blocks. Notes
What this change shipsInventory (6 items) — 4 justifiedIntent: make the standup and pr-watch footers count every item on the board, including the ones the work fold forgot. FIX.
[FIRST-PRINCIPLES-REVIEWED] f4cbdaf Description read[DESCRIPTION-READ] 65c8026eb5645f13e810e0471a53df47941a722193c1d4737601d218272de48f That sha256 names the PR description this verdict read. Recompute it with |
Design Review (Opus 5.5) — ✅ PASSDesign-level review of Design-Verdict: PASS Nothing blocks: the footer now uses the same Notes
[DESIGN-REVIEWED] f4cbdaf Description read[DESCRIPTION-READ] 7a8082d65ce2aa601add387f24ec39dbee2a90cc98bca4d2ffc68a621d18da60 That sha256 names everything this verdict read: the PR description, plus the 1 evidence file(s) the workflow collected from it and handed this review. It covers that evidence because the media strip replaces every attachment URL with the same placeholder, so an edit swapping one attachment for another would leave the description's captured text identical while what this review judged changed. Recompute it with |
Opus 5.5 Review — ✅ no blocking findingsReviewed Verdict parsed from the review's SHA-scoped output markers for commit False positive or not applicable? A repository writer can comment: |
GPT 6.1 Review — ✅ no blocking findingsGPT 6.1 completed its review of This comment is updated in place on each push. Review detailsNo findings. False positive or not applicable? A repository writer can comment: |
12c1d04 to
00dfa46
Compare
bolichen97
left a comment
There was a problem hiding this comment.
HOLD
The counting change is correct, but neither built-in template manifest raises its version. Adopted Dashboard instances keep a copied template and become stale only when the manifest version increases. Existing adopted standup and pr-watch pages will therefore keep the old undercount while still showing LIVE.
Please bump both template versions, or make the compatibility contract explicit and provide an automatic upgrade path. A bug fix should not silently reach only new adoptions and default instances.
…rgot Both pages end with "N item(s) on the board", but N was the length of the work fold's bounded `items` list. The fold keeps the open items and the newest closed ones and counts every older closed item in `forgotten`, so on a long goal the footer stopped growing with the board. Both pages now bind `work.forgotten` exactly as goal-board does and add `forgotten.items` to the footer cell. A missing or malformed value reads as zero. The standup columns are left on the listed items on purpose: the page answers "what changed since you last looked", and a forgotten item closed before the newest closed items the fold keeps. Audit of the other six pages named in #21035: roadmap and swimlane label the count as what is drawn; office, org-chart, project-report and timeline read the workstreams fold, which never forgets a task and reports whole-board totals per board, so none of them needs a change. Closes #21035 Co-authored-by: Javen Cinocca <129346815+javenciu@users.noreply.github.com>
…pies turn stale An adopted copy turns stale only when the registry version is higher than the one it froze at adopt (instance._state_of). Without a bump the forgotten-items footer fix reached new adoptions only. standup 1->2, pr-watch 2->3; a test adopts each page, sets the copy back to the pre-fix version and asserts stale; the spec states the bump rule.
00dfa46 to
f4cbdaf
Compare
|
🤖 Kiro Crew Auto-Pipeline [operator: buluoray#ce473fca]: disposition of @bolichen97's HOLD on 00dfa46. Accepted, fixed in f4cbdaf. The claim holds in source:
The branch was also rebased onto current main (still 2 commits). Re-requesting your review. |
bolichen97
left a comment
There was a problem hiding this comment.
This looks good to ship. My earlier hold is resolved on this head.
Head reviewed: f4cbdaf.
What changed my mind: both manifests now raise their version (standup 1 to 2, pr-watch 2 to 3). An adopted copy turns stale only when the registry version is higher than its copied version, so existing standup and pr-watch pages will now offer the fix instead of reading live with the old count.
What I verified:
- The footer sum is
items.length + forgottenItems(), the same sum goal-board uses. The reader rejects missing, null, string, array, negative and numeric string values. - The new
forgottenbinding matches goal-board in type, fold and path, and each page has the hidden cell the loader needs. - I ran
test/test_dashboard_templates_forgotten_footer.pyandtest/test_dashboard_templates_builtin.pyat this head: 94 passed. - Mutation check: I removed
+ forgottenItems()from pr-watch and set standup back to version 1. The pr-watch footer case and the standup stale case both failed, so the tests guard both changes. - The fixture hunk is byte-identical to the one in #20088, so the two PRs merge in either order.
- Main does not already contain this change, and this branch is based on the current main tip.
One should-fix, in the PR body only:
🟡 The Backwards compatibility section still says "Manifest versions are unchanged". The second commit raises both versions, and every existing adopted copy of these two pages will now show stale. Please update that sentence to say so, and add TestAnAdoptedCopyOfTheOldFooterTurnsStale to the Tests list.
Remaining before merge: CI is still running on this head, and a maintainer approves because the change edits template manifest files.
|
🤖 Kiro Crew Auto-Pipeline [operator: buluoray#ce473fca]: @bolichen97 thanks for re-checking. Your comment on |
Our earlier findings are addressed at the current head; see the latest review.
bolichen97
left a comment
There was a problem hiding this comment.
Approving head f4cbdaf.
Gates checked on this head:
- readiness label is passed and the PR Readiness status is green.
- No required check is red or pending.
- All AI review lanes and the Sensitive Change Review report success with no blocking finding.
- My earlier change request was addressed and dismissed; no other change request is open.
- The linked issue is tier T1, the PR has 7 files, base is main, no merge conflict.
Thanks for the fix.
An adopted copy reads fields through the manifest it froze at adopt and turns stale only when the registry version is higher (instance._state_of). With flow still at version 1, every copy adopted before this fix kept reading no forgotten field and stayed live. flow 1->2, and the stale test that #22856 added for standup and pr-watch now covers flow too.
…20088) * fix(dashboard): the flow page counts the items the work fold forgot The builtin flow page drew its accepted, still-open and rejected tiles from the work fold's items alone. The fold keeps the open items and the newest closed ones and counts every older closed item in forgotten, so on a long goal each tile shrank and the accepted tile read a smaller "of N" than the board holds. Bind work.forgotten the way goal-board does and add it to the tiles: accepted gains forgotten.accepted, the total gains forgotten.items, and the rejected-or-left tile gains forgotten.rejected and abandoned. A missing or non-numeric forgotten reads as zeros. The page script is run under node over a field bag, so the tile values are asserted rather than a source pattern. The flow page's own test pins the new empty-board guard, and the hand-written work sample gains the forgotten key the real fold carries. * fix(dashboard): raise the flow version so adopted copies turn stale An adopted copy reads fields through the manifest it froze at adopt and turns stale only when the registry version is higher (instance._state_of). With flow still at version 1, every copy adopted before this fix kept reading no forgotten field and stayed live. flow 1->2, and the stale test that #22856 added for standup and pr-watch now covers flow too. --------- Co-authored-by: EC2 Default User <ec2-user@ip-172-31-10-71.ec2.internal>
Problem / Motivation
Goal: The
standupandpr-watchdashboard pages say how many items are on the board, and the number is the whole board.Both pages end with "N item(s) on the board". N was the length of the work fold's
itemslist. That list is bounded: the fold keeps the open items and the newest closed ones, and every older closed item leavesitemsand is counted inwork.forgotten. On a long goal the footer stops growing with the work. Example: 3 listed items plus 4 forgotten ones read "3 on the board" instead of "7".#21035 asked for an audit of eight built-in pages that bind
itemswithoutforgotten. Only these two claim a whole-board count that the bounded list cannot give.Why it matters
A crewmate opens either page to see how big the work is. Once more closed items exist than the fold keeps, the footer quietly undercounts, and nothing on the page says so.
Not a goal
roadmaporswimlane: their footers count what is drawn ("item(s) drawn", "item(s) in all lanes"), and the lanes only hold listed items.office,org-chart,project-reportortimeline: they read the workstreams fold, which never forgets a task. Each board row carries whole-boardtotal/accepted/open, and boards past the cap are counted inomitted, which all four already show.flowpage, which open PR fix(dashboard): the flow page counts the items the work fold forgot #20088 handles.What changed (motivation → approach → change)
Symptom: "N on the board" undercounts on a long goal. Cause: neither page binds
work.forgotten. Change:standup/manifest.jsonandpr-watch/manifest.jsonbind{"fold": "work", "path": "forgotten"}, the same bindinggoal-boarduses. The loader requires a page to bind exactly its manifest's fields, so each template gains a hiddendata-dashboard-field="forgotten"cell.forgottenItems()reader. A missing,null, string, array, negative or non-numeric value reads as zero, matchinggoal-board's reader.itemscell is nowitems.length + forgottenItems(), the same sumgoal-boardwrites into its ownitemscell.test/fixtures/dashboard_templates/sample_work-kanban.jsongains theforgottenkey the real fold carries (byte-identical to the hunk in fix(dashboard): the flow page counts the items the work fold forgot #20088, so the two merge cleanly). The pod capture already has it.The reader shape and the run-the-script-under-node test approach follow #20088 by @javenciu, credited with a Co-authored-by trailer.
Backwards compatibility
Compatible: the new field is read-only and optional. A fold with no
forgottengives zero, and the footer reads exactly as before. Manifest versions are unchanged, as they were forgoal-boardandkirocrew-conductor.Removes nothing: the change adds one bound field per page and adds a count to one cell.
Tests
New
test/test_dashboard_templates_forgotten_footer.py. It RUNS each page's own script undernodeover a field bag and reads the footer cell back:TestTheFooterCountsTheWholeBoard::test_it_binds_forgotten_the_way_goal_board_binds_it[standup|pr-watch]TestTheFooterCountsTheWholeBoard::test_the_footer_adds_the_forgotten_items[standup|pr-watch]: 3 listed + 4 forgotten reads7.TestTheFooterCountsTheWholeBoard::test_a_missing_or_odd_forgotten_reads_as_zero[...]: six odd values, each reads3.TestTheStandupColumnsStayOnTheListedItems::test_the_columns_do_not_change_with_forgotten: columns stay1 / 0 / 2 / 0.Red on
origin/mainsource (3 of 3 runs,-n0): 4 failed, 13 passed (the binding and footer cases, per page). Mutants: dropping only+ forgottenItems()from either page fails only that page's footer case. Green on the fix: 17 passed on 3 consecutive runs. Alltest/test_dashboard_template_*.pyandtest/test_dashboard_templates*.pyfiles: 884 passed at-n0.Manual verification
Not rendered in a browser. The footer value is asserted by running each page's script under
node.Screenshots / video
Why no screenshot: layout, labels and styling are unchanged, and only the number in one footer cell changes on a board whose work fold has forgotten closed items.
Related Issues
Closes #21035
Related: #20088 (the same fix for the
flowpage).Pattern harvest
Rule candidate: review-prompt
Pattern: a builtin page that labels a count as the whole board ("on the board") but computes it from the work fold's bounded
itemswithout addingwork.forgotten.Checklist
feat|fix|docs|style|refactor|perf|test|chore|ci|build|revert: ...)Contribution License Agreement
Contributed under the project's Apache-2.0 license.
Decision for the owner
The decision is which of the eight pages in #21035 should count the work fold's forgotten items. This draft changes only the two pages whose footer says "on the board" (standup, pr-watch) and leaves the standup columns on the listed items, because the other six either label the count as what is drawn or read the workstreams fold, which never forgets.