Repository navigation
finding(plugin-detail): on the record page's discussion, a stored reply never shows and the reply input is never drawn, because nothing sets replyCount on read #12144
Description
Filing gate ①: a product defect with a named producer. Filed by the domain:ui seat 2 (seat post #9771, session_01TYgwmFK1q4KJ6Qq2WRLzsD) from an out-of-scope finding in the objectui#12082 round-2 dev report (PR #12140, report comment 6109609151). The seat re-read the code below before filing. ⛔ Not a claim.
Read on objectui 6dea96f (PR #12140's head; these lines are unchanged from main)
-
reach:named real producer. The synthesized default record page composesrecord:discussion(packages/plugin-detail/src/synth/buildDefaultPageSchema.tsabout:969). That block renders throughpackages/plugin-detail/src/renderers/record-chatter.tsx, whose config setsfeed.enableThreading: true(about:172). So threading is on for every record's default page. -
With threading on,
packages/plugin-detail/src/RecordActivityTimeline.tsxdoes two things:- it removes every item with a
parentIdfrom the top-level list (about:298); - it draws
ThreadedRepliesonly whenitem.replyCount > 0(about:634). That component is the only place both the reply list and the reply input (showReplyInput={!!onAddReply}) render.
- it removes every item with a
-
packages/app-shell/src/views/RecordDetailView.tsxmaps eachsys_commentrow read to aFeedItemwithparentIdbut neverreplyCount(about:2113). OnlyhandleAddReplybumpsreplyCount, after a reply is posted (about:2347), and posting needs the reply input. -
Result, from the code read:
- a reply stored on the server is shown neither at the top level nor under its parent;
- no comment read from the server offers a reply input, so the first reply on a comment cannot be posted from the page.
NOT MEASURED at a running page: this is a code read.
Done when
- On the default record page, a stored reply is shown under its parent after a reload.
- A comment with no replies offers the reply input wherever the host hands over
onAddReply. - Pins, through the record page's feed:
- a
sys_commentread with a parent and one child shows the child under the parent; - a parent with no child offers the reply input;
- CONTROL: with threading off, the list is unchanged.
- a
Who acts, and in what order
- The
domain:uiexecution seat, after triage grades this card. - Serial: after PR fix(core,app-shell,plugin-detail): freeze the write census's backlog; the record feed and a related list's Add existing / fallback remove read the affordance-to-grant map (objectui#12082) #12140 lands. It edits
RecordDetailView.tsx's feed handlers.
Dedupe
Two repo-scoped issue searches on objectui, open and closed:
- "record feed threaded replies hidden after reload replyCount never derived reply input unreachable": 0 hits.
- "comment reply parent_id reply count discussion thread not shown": 1 hit, objectui#875 (closed; the discussion panel absent from a drawer, a different defect).
Activity
objectstack-fleet commentedon Oct 11, 2026
Triage: first grade,
bug·priority:p2·domain:ui·area:records, intopm:queue. Measure firstTriage seat (seat post objectstack-ai/objectstack#6015) ·
session_01AavokzJ5DndAwitDXvKy4U· 2026-10-11T13:58Z. ⛔ Not a claim, ⛔ not a dispatch.RecordDetailView.tsx(packages/app-shell) andRecordActivityTimeline.tsx(packages/plugin-detail) aredomain:ui.replyCounton read, in the same placeparentIdis mapped (RecordDetailView.tsx'ssys_comment→FeedItem), by counting children in the read set.ThreadedRepliesdraws, and the reply input with it, for any comment wherever the host passesonAddReply.replyCountgate.Clause-②: no.objectstack-fleet commentedon Oct 11, 2026
Claim: PM loop round 1 (measure both effects first, then derive
replyCounton read, as triage6109788149directs)Session:
session_016djJF12Qt14ejKR5Vjt4sKAccount:
os-steveBranch:
claude/issue-12144-reply-count-on-readWorktree:
objectui-issue-12144Domain:
domain:uiSeat:
domain:ui#1File surface:
packages/app-shell/src/views/RecordDetailView.tsx(thesys_commentrow toFeedItemread mapping only);packages/plugin-detail/src/RecordActivityTimeline.tsx(only the zero-reply input condition, keeping its onereplyCountgate, and only if the measurement shows it is needed); new pins beside each; the docs page that describes the feed's threading, if one says otherwise;.changeset/12144-reply-count-on-read.md(stop on breach; explain in the report)Container & model:
M,mode:subagent,model: default tier — dispatch-gates --tier --repo objectstack-ai/objectui: "no path-derived mandate"Clause-②: no
Responsibility:
objectui's record page read mapping: RecordDetailView.tsx maps each sys_comment row to a FeedItem with parentId and never replyCount, so the timeline's replyCount gate hides stored replies and the reply input | none: the timeline's one replyCount gate already exists and is only never fed on read | every user of a default record page, where record-chatter.tsx sets enableThreading: true; nobody can post a first reply from a comment read from the server todayThread-read: 6109788149
Serial constraints cleared: none blocking. Read at 2026-10-11T14:48Z on objectui
main158c75d:RecordDetailView.tsx's feed handlers, landed as158c75d. The dev starts from that merged code.RecordDetailView.tsxread-only; its branch head50c7747changes 28 files, neither file above among them. objectui#12112 and console: the Studio app's own root still rendersStudioHomePage, the page the maintainer calls retired. Decide whether to retire it, and what the root shows instead #12113 touch neither file.RecordDetailView'ssys_permission_setslot, a different region from the feed mapping.Clause-②: noholds only if no prop, export or locale key is added. The dispatch forbids adding one without stopping to report, so the declaration does not depend on a route the dev has not chosen.objectstack-fleet commentedon Oct 11, 2026
os-dev-report
{ "issue": 12144, "status": "done", "branch": "claude/issue-12144-reply-count-on-read", "pr": "https://github.com/objectstack-ai/objectui/pull/12147", "session": "session_016djJF12Qt14ejKR5Vjt4sK (subagent: the dispatching seat's id, the Claude-Session line of this run)", "premise_still_valid": true, "summary": "Measured first on 158c75d through the synthesized default record page (RecordDetailView, no authored page, real sys_comment read and mapping, real panel): a stored reply showed nowhere and no comment read from the server offered the reply input (new page pin 6 red / 1 green, the green being the threading-off control). RecordDetailView now counts each comment's parent_id children among the rows its sys_comment read returned and sets replyCount in the same map step as parentId, which is what objectstack's sys_comment object definition prescribes (count parent_id children at read time). RecordActivityTimeline offers the reply input on every comment row the host hands onAddReply for, replies or not, and on no non-comment row (module-private canTakeReply, the objectui#11035 reasoning); the reply list keeps its one replyCount gate. No prop, export, type member or locale key added, so Clause-②: no holds. H1-H4 all confirmed; the feedComment grant behaviour from objectui#12140 is unchanged and pinned both ways.", "tests": "BASE (7a19eaa = base + pin only): pnpm exec vitest run packages/app-shell/src/views/RecordDetailView.replyCountOnRead-12144.test.tsx → 'Tests 6 failed | 1 passed (7)' (first failure: Unable to find an accessible element with the role \"button\" and name /1 reply/, after the child-absent assertion passed; input cases: Unable to find an element with the placeholder text of: Reply…). FIX: both new pins 'Tests 13 passed (13)'. ABLATIONS via objectstack scripts/ablation-replace.mjs, each 'ok mutation landed' (anchor 1 → 0, blob changed) and 'ok restored: blob == HEAD and git diff HEAD is empty'; vitest resolves @object-ui/plugin-detail to src via the root alias table, no dist in the path: (1) replyCount: undefined in the read → 3 failed | 4 passed (the list cases); (2) input re-gated on replyCount → 6 failed | 7 passed (the three input cases in each file); (3) input on every row → 2 failed | 11 passed (the activity-row cases); (4) list without the replyCount gate → 1 failed | 12 passed. SUITES at f7f477a: vitest run packages/plugin-detail/ → 'Test Files 256 passed | 1 skipped (257)', 'Tests 2491 passed | 8 skipped'; vitest run packages/app-shell/ → 'Test Files 1270 passed | 1 skipped (1271)', 'Tests 12306 passed | 9 skipped'; os-verify-lock 'VERDICT command-exit 0'. type-check (hyphenated, script name echoed) for both packages: Done, VERDICT command-exit 0; tsc -p tsconfig.test.json --listFiles lists both new pins (1 hit each, control 1 hit).", "mcp_calls": "0 — no MCP tool used (card read through the public-repo payload channel, the issue page's embedded JSON: timelineItems totalCount 12 = edges 12, hasNextPage false)", "api_writes": "2 relay strokes, each one POST /repos/objectstack-ai/objectstack/dispatches executed as objectstack-fleet[bot]: (1) pr_create = POST /repos/objectstack-ai/objectui/pulls (draft) + POST /repos/objectstack-ai/objectui/issues/12147/assignees (os-steve), read-back 9233 bytes sent = stored; (2) comment = POST /repos/objectstack-ai/objectui/issues/12144/comments (this report). git push (not REST): 4 pushes of claude/issue-12144-reply-count-on-read (empty-branch probe, pin, fix, changeset). Labels: zero writes.", "open_questions": [], "out_of_scope_findings": [ "carrier: PR objectui#12147 (a round-2 patch, if the seat widens the claim's file surface by one test file) · noted, not filed — the header of packages/app-shell/src/views/RecordDetailView.commentWriteFailure-10899.test.tsx says the reply writer's input 'is not reachable from a loaded feed … no read sets that'; this change makes it false, and the reply writer's failure path is now reachable through the page. Outside the claim's surface, so not edited (stop-on-breach). Dedupe words: replyCount, commentWriteFailure, handleAddReply, reachable", "carrier: 承接者:无 · noted in the PR's Acceptance notes only — content/docs/plugins/plugin-detail.mdx 'Console Integration' sample uses camelCase sys_comment fields (threadId, _id, author) and the ThreadedReplies table gives showReplyInput's default as false where the code defaults to true; unexercised docs drift, no reach measured", "carrier: 承接者:无 · noted in the PR's Acceptance notes only — record:discussion's page window (feed.limit, default 20) counts replies; a reply whose parent is older than the window shows nowhere until Load more reaches the parent. Pre-existing paging, read not measured at a public door", "carrier: 承接者:无 · noted in the PR's Acceptance notes only — the sys_comment read carries no $top, so a reply whose parent is outside the server's default page counts toward no item and stays hidden with threading on; no second fetch invented, as the dispatch asked" ], "gates": [ "① deps build: pnpm exec turbo run build --filter='@object-ui/app-shell^...' --filter='@object-ui/plugin-detail^...' --concurrency=2 → 'Tasks: 28 successful, 28 total', VERDICT command-exit 0", "② type-check: pnpm --workspace-concurrency=2 --filter @object-ui/plugin-detail --filter @object-ui/app-shell run type-check → both 'type-check: Done', VERDICT command-exit 0", "② tests: vitest run packages/plugin-detail/ (256 files passed, 1 skipped) and packages/app-shell/ (1270 files passed, 1 skipped), VERDICT command-exit 0 at f7f477a", "③ node scripts/check-changeset-presence.mjs → exit 0 ('4 source file(s) of 2 released package(s) changed, and this change declares 1 changeset(s)')", "③ pnpm check:changeset-claims → exit 0 ('No pending changeset names a file this change touches')", "③ pnpm check:pending-changeset-literals → exit 0", "③ pnpm check:control-bytes → exit 0 ('OK (scanned 8371 tracked text file(s)')", "③ pnpm check:new-line-citations → exit 0 ('VERDICT new-cross-file-line-citations: 0 new citation(s)')", "③ pnpm check:phantom-deps → exit 0", "docs gates (docs:check-links, check:doc-fences, check:doc-json): not owed, no docs touched", "governed: node scripts/check-governed-queue-guard.mjs --test on the 5 paths → NOT GOVERNED", "lint (narrowed, declared): eslint --format json on the 4 changed .ts/.tsx files → 4 files, 0 errors; RecordDetailView.tsx 109 warnings at base and head; no parserOptions in eslint.config.js (no type-aware linting) and no rule under eslint-rules/ reads another file (grep zero, control grep hit 1), so untouched files' verdicts cannot move; repo-wide lint left to CI", "browser: NOT MEASURED, reason: ../objectstack in this container has no node_modules and no built spec/cli dist, so the per-agent backend cannot boot inside the foreground budget", "CI: not awaited (in_progress at report time; PM owns convergence)" ], "line_budget": "n/a — no skills/** or line-ratcheted ledger touched; diff 5 files, +479 / -3", "deviations": [ "Commit trailers use the AGENTS.md model-free pair (Claude-Session + Co-authored-by: Claude) instead of the harness reminder's model-named Co-Authored-By, and the PR footer is os-dev's session-URL form instead of the harness's other footer shape; the harness reminder yields to the repo/contract rule by its own precedence sentence", "Lint is the declared narrowing above, not the repo-wide run", "The timeline's input predicate is unified (onAddReply and a comment row) for every row, so a non-comment row with replyCount above zero no longer offers the input; no producer sets replyCount on a non-comment row (the record page counts only sys_comment rows), stated in the PR", "The app-shell suite (1271 files) held the verify lock for 54 min; a sibling objectui#12145 run queued behind it during that window" ], "files_changed": [ ".changeset/12144-reply-count-on-read.md", "packages/app-shell/src/views/RecordDetailView.tsx", "packages/app-shell/src/views/RecordDetailView.replyCountOnRead-12144.test.tsx", "packages/plugin-detail/src/RecordActivityTimeline.tsx", "packages/plugin-detail/src/__tests__/RecordActivityTimeline.zeroReplyInput-12144.test.tsx" ] }Generated by Claude Code
objectstack-fleet commentedon Oct 11, 2026
Patch round 1 on PR objectui#12147 (head
f7f477a), with a file-surface amendment ·domain:uiseat 1 ·session_016djJF12Qt14ejKR5Vjt4sK· 2026-10-11T16:08ZNot a verdict yet. Report
os-dev-report6110947036read. The diff was reviewed against GitHub, and it holds:parent_idand looked up byid, the timeline's own grouping;onAddReplyis still handed over only underfeedGates.comment(RecordDetailView.tsx:3257onmain);sys_comment's docblock (objectstackpackages/plugins/plugin-audit/src/objects/sys-comment.object.ts:37) says to countparent_idchildren at read time;Claim amendment (to
6110239138): the file surface gains one test file,packages/app-shell/src/views/RecordDetailView.commentWriteFailure-10899.test.tsx.Clause-②: nois unchanged.objectstack-fleet commentedon Oct 11, 2026
os-dev-report
{ "issue": 12144, "status": "done", "branch": "claude/issue-12144-reply-count-on-read", "pr": "https://github.com/objectstack-ai/objectui/pull/12147", "head": "dc1f3c0 (git rev-parse --short HEAD after the last commit; pushed, remote branch head dc1f3c0c50154da6c8ad0fb1fe33d01236e0d37c)", "session": "session_016djJF12Qt14ejKR5Vjt4sK (subagent: the dispatching seat's id, the Claude-Session line of this run)", "premise_still_valid": true, "summary": "Patch round 1, under the claim amendment in card comment 6110965484, which I read on the card. The header of RecordDetailView.commentWriteFailure-10899.test.tsx now says what is true: the reply writer's input is reachable from a loaded feed, because the timeline offers it on every comment row the page hands onAddReply for (objectui#12144). The plugin-detail draft pin is still named as the input's own half. The file gains one case that pins handleAddReply's failure path in the same shape as the comment writer's case. The page reads one sys_comment row with no replies, and the case types into that comment's reply input and sends. create rejects with the file's default 404. The case asserts four things: toast.error is called once with a message that does not contain sys_comment; create was called with thread_id, parent_id c1 and the body; no reply row and no reply-count toggle appear; the draft stays in the input. main has not moved from 158c75d, so there was no merge; the patch is pushed. Nothing else changed: no prop, export or locale key, and Clause-②: no stands.", "tests": "At dc1f3c0. The amended file alone: 'Tests 4 passed (4)'. ABLATION with the round-1 ablation-2 anchor, through objectstack scripts/ablation-replace.mjs on packages/plugin-detail/src/RecordActivityTimeline.tsx, replacing 'const offersReply = !!onAddReply && canTakeReply(item);' with 'const offersReply = !!onAddReply && (item.replyCount ?? 0) > 0;'. The tool printed 'ablation-replace: ok mutation landed: anchor 1 -> 0, blob 4f8f20e78f47 -> a010fdf85c64'. The new case went red with 'Unable to find an element with the placeholder text of: Reply…', 'Tests 1 failed | 3 passed (4)'. Restore line: 'ablation-replace: ok restored: blob == HEAD (4f8f20e78f47) and `git diff HEAD` is empty'. RE-RUN of the two new pins plus the 10899 file: 'Test Files 3 passed (3)', 'Tests 17 passed (17)', VERDICT command-exit 0. type-check for @object-ui/app-shell echoed '> tsc --noEmit && tsc -p tsconfig.test.json', VERDICT command-exit 0. The dependency closure came back as a turbo cache replay ('28 cached, 28 total'), and the replayed plugin-detail dist carries the fix: it contains the offersReply and replyCount-gated replies expression. Per the dispatch, the full app-shell suite was not re-run; CI runs it.", "mcp_calls": "0. No MCP tool was used. I read the card through the public-repo payload channel: timeline totalCount 14 = edges 14, hasNextPage false; comment 6110965484 is present.", "api_writes": "1 relay stroke: POST /repos/objectstack-ai/objectstack/dispatches, executed as objectstack-fleet[bot]. Its one comment action is POST /repos/objectstack-ai/objectui/issues/12144/comments (this round-2 report). git push is not REST: 1 push (dc1f3c0). No PR-body PATCH. Labels: zero writes.", "open_questions": [], "out_of_scope_findings": [ "carrier: 承接者:无 · noted in the PR's Acceptance notes only. content/docs/plugins/plugin-detail.mdx: the 'Console Integration' sample uses camelCase sys_comment fields, and the ThreadedReplies table gives showReplyInput's default as false where the code defaults to true. Docs drift nothing exercises; unchanged from round 1", "carrier: 承接者:无 · noted in the PR's Acceptance notes only. record:discussion's page window counts replies, and a reply whose parent is older than the window shows nowhere until Load more reaches the parent. Unchanged from round 1", "carrier: 承接者:无 · noted in the PR's Acceptance notes only. The sys_comment read carries no $top, so a reply whose parent is outside the server's default page stays hidden. No second fetch. Unchanged from round 1" ], "pr_body_edits_for_the_seat": [ "Under '## Acceptance notes', replace the bullet that starts '**A header comment this change falsifies.**' with: '**The reply writer, pinned through the page (patch round 1).** The header of `RecordDetailView.commentWriteFailure-10899.test.tsx` now says the reply input is reachable from a loaded feed. The file pins `handleAddReply`'s failure path through the reply input under a comment read from `sys_comment`: create rejects, one localized error, no reply row, the draft kept. With the input gated on `replyCount` again, the case is red (`Unable to find an element with the placeholder text of: Reply…`, `1 failed | 3 passed`), and the mutation was restored to blob equal to HEAD.'", "Under '## Pins', add a third bullet: '`packages/app-shell/src/views/RecordDetailView.commentWriteFailure-10899.test.tsx` (existing file, header corrected), one case added: a rejected reply create under a comment read from the server adds no reply, raises a localized error and keeps the draft.'", "The heading '## Verification (all on `f7f477a`; ...)' gets an added line: 'Patch round 1 at `dc1f3c0`: the two new pins and the 10899 file 17/17 green, and `@object-ui/app-shell` type-check green. The full app-shell suite was left to CI.'" ], "gates": [ "pins: vitest run of RecordDetailView.replyCountOnRead-12144.test.tsx, RecordActivityTimeline.zeroReplyInput-12144.test.tsx and RecordDetailView.commentWriteFailure-10899.test.tsx → 'Test Files 3 passed (3)', 'Tests 17 passed (17)', VERDICT command-exit 0 at dc1f3c0", "ablation (ablation-2 anchor) on the new 10899 case → 'Tests 1 failed | 3 passed (4)'; restored, blob == HEAD (4f8f20e78f47), git diff HEAD empty", "deps: pnpm exec turbo run build --filter='@object-ui/app-shell^...' --concurrency=2 → 'Tasks: 28 successful, 28 total', 'Cached: 28 cached, 28 total', VERDICT command-exit 0. The cached plugin-detail dist carries the fix (grep of dist/index.js)", "type-check: pnpm --filter @object-ui/app-shell run type-check → '> tsc --noEmit && tsc -p tsconfig.test.json', VERDICT command-exit 0 at dc1f3c0", "node scripts/check-changeset-presence.mjs → exit 0 ('5 source file(s) of 2 released package(s) changed, and this change declares 1 changeset(s)')", "pnpm check:changeset-claims → exit 0; pnpm check:pending-changeset-literals → exit 0; pnpm check:control-bytes → exit 0; pnpm check:new-line-citations → exit 0 ('0 new citation(s)')", "lint, narrowed: eslint --format json on the edited test file → 1 file, 0 errors, 4 warnings at both f7f477a and dc1f3c0 (no type-aware linting, as in round 1)", "full app-shell and plugin-detail suites: not re-run this round, per the dispatch (CI runs them). Round 1's runs at f7f477a stand for the unchanged source", "CI: not awaited (in_progress at report time; CI convergence belongs to the PM)" ], "line_budget": "n/a. No skills/** file or line-ratcheted ledger touched. The round adds one file to the diff (+45 / -5); the PR is now 6 files", "deviations": [ "The worktree was removed after round 1 per the cleanup rule. I recreated it on the existing branch at f7f477a (local equal to remote), ran pnpm install again, and removed it again after this round", "Commit trailers are the AGENTS.md model-free pair, not the harness reminder's model-named form (as in round 1)", "Lint is the declared narrowing, not the repo-wide run", "Lock waits this round: 24s, 141s and 173s while sibling runs held it; the narrow runs each held it 1s to 102s" ], "files_changed": [ ".changeset/12144-reply-count-on-read.md", "packages/app-shell/src/views/RecordDetailView.tsx", "packages/app-shell/src/views/RecordDetailView.replyCountOnRead-12144.test.tsx", "packages/app-shell/src/views/RecordDetailView.commentWriteFailure-10899.test.tsx", "packages/plugin-detail/src/RecordActivityTimeline.tsx", "packages/plugin-detail/src/__tests__/RecordActivityTimeline.zeroReplyInput-12144.test.tsx" ] }Generated by Claude Code
objectstack-fleet commentedon Oct 11, 2026
ACCEPT: PR objectui#12147 (head
dc1f3c0c50), round 1 plus patch round 1 ·domain:uiseat 1 ·session_016djJF12Qt14ejKR5Vjt4sK· 2026-10-11T16:22ZReviewed against GitHub and the diff, not the reports (
os-dev-report6110947036and round-26111079611).Verdict
main. The first line isFixes #12144, andClause-②: nosits at a line start.os-steve.6110965484. Nothing undercontent/docs/releases/.check-governed-merges --pr objectstack-ai/objectui#12147: NOT governed, 532 changed lines.RecordDetailView.tsxcounts each row'sparent_idamong the rows thesys_commentread returned, and setsreplyCountfrom that count. It is keyed the way the timeline groups replies (repliesByParent.get(item.id)).RecordActivityTimeline.tsxkeeps onereplyCountgate for the reply list. It offers the input on every comment row handedonAddReply(canTakeReplyis module-private).feedGates.comment(RecordDetailView.tsx:3257onmain).Clause-②: noholds: the diff adds no prop, export, type member or locale key.patch. The two bullets match the two hunks.parent_idchildren at read time": read at objectstackpackages/plugins/plugin-audit/src/objects/sys-comment.object.ts:37.sys_comment": matchesfeedGates.comment.1 failed | 3 passed), and every restore was blob-equal to HEAD.replyCountabove zero no longer offers the input. Onmain, the only producer ofreplyCountoutside tests is the record page's comment mapping (RecordDetailView.tsx:2356).dc1f3c0(read at this record): 29 success, 3 skipped by design, 10 in progress (Type Check, the eightTestshards,Spec Main Shape Gate). The redTestonf7f477awas the eight shards cancelled by the patch push. That head is superseded and the reading is not counted.Out-of-scope findings
10899test header made false by this change: fixed in patch round 1 (claim amendment6110965484), with a case that pins the reply writer's failure path through the page.plugin-detail.mdxsample fields and theshowReplyInputdefault: Acceptance notes. This is docs drift that nothing exercises.Acceptance notes
dc1f3c0, and its Verification section predates that head. The final readings are in6111079611and above.Landing: ready and auto-merge through the merge queue once every check on
dc1f3c0issuccessor skipped by design.objectstack-fleet commentedon Oct 11, 2026
Landed: PR objectui#12147 →
maina7d4920·domain:uiseat 1 ·session_016djJF12Qt14ejKR5Vjt4sK· 2026-10-11T16:50Z158c75d. The card closed ascompletedthroughFixes #12144.dc1f3c0c50. ACCEPT6111097861covers that head. There is no contract review, because the claim isClause-②: noand the diff touches no contract face.mainnow:onAddReplyoffers the reply input, still under thefeedCommentgrant.pm:dispatchedis removed in this act.domain:ui,area:recordsand the grade stay.