Repository navigation
console: record comments cannot be edited, deleted or replied to; record:activity declares enableThreading: true and no reply control renders (objectui#12211 item 5) #12249
Description
Filing gate ①: a reproduced product defect, split from list card #12211 (item 5) as triage 6121957396 directs ("The dispatching seat splits it per item into sub-issues"). This card carries item 5 only. With it every item of the parent is split, and the parent closes when every item is closed.
Parent: #12211. Domain and priority are inherited from the parent (domain:ui, priority:p3); area:records per triage.
Reader: the domain:ui lane (either seat), from pm:queue.
Dedupe: the parent's table, item 5: comment edit delete reply (127 hits, none on point in the first page).
Triage's note for this item (6121957396), verbatim: "Item 5: the dev first reads what the spec declares for comment edit, delete and reply. enableThreading: true with no reply is a declared prop not honoured, so it is in scope. An edit or delete capability nothing declares is a new card, not this list."
What happens (from the parent, item 5)
Source: the 2026-10-12 hosted-shape browser pass of HotCRM 4.0.1 on a cloud rig at production v1.36.0's pins (objectui d758f2f20d), zh-CN.
- Reach: CASE-00041 (活动) and KA-0005 (讨论). Measured 2 of 2.
- After
POST /api/v1/data/sys_comment→ 201, the only control on the comment is 添加表情回应 (add reaction), plus 查看来源 (view source) on some entries. case_detail_page'srecord:activitysetsenableThreading: true, yet no reply control renders.
- After
Acceptance
- Replies render where threading is enabled (the declared
enableThreading: trueis honoured), or the prop's declaration is corrected if threading is not supported. - An author can edit and delete their own comment only if the spec declares that capability. If nothing declares it, that half is a new card, not this one (triage's note above).
Generated by Claude Code
Activity
objectstack-fleet commentedon Oct 12, 2026
Claim: PM loop round 1 (measurement first: no fix is written before the measurement below exists)
Session:
session_01CsyMRsFJBKNEMa4oPY1b5zAccount:
zhuangjianguo(the seat's linked user asGET /useranswers it; the card's assignee)Branch:
claude/issue-12249-activity-threading-replyWorktree:
objectui-issue-12249Domain:
domain:uiSeat:
domain:ui#1File surface: an authored
record:activity's reply path, read on objectuimain0cd322dat 2026-10-12T14:36Z (stop on breach; explain in the report):packages/plugin-detail/src/renderers/record-activity.tsx, which readsenableThreadingand takes its write handlers (onAddReplyamong them) off the hostDiscussionContext.packages/plugin-detail/src/RecordActivityTimeline.tsx, whose reply list and reply input render underenableThreading(objectui#12144'scanTakeReplygate, PR objectui#12147)..changeset/12249-activity-threading-reply.mdif a fix ships.packages/app-shell/src/views/RecordDetailView.tsx(theDiscussionContextproducer and thesys_commentread). objectui#12244 of this seat holds that file; a fix that needs it stops and reports.RecordActivityPropsdeclares none (spec@objectstack/spec@17.7.0, objectui's resolved version), so that half is a new card per triage6121957396, not this one.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: an authored record:activity with enableThreading true drew no reply control at HotCRM's pin d758f2f20d | the platform path is objectui#12144's fix (PR objectui#12147, merge a7d4920), which is on main and 34 commits after that pin, measured through the default record:discussion page only | any user on a record page whose author declared record:activity with enableThreading; measured on hosted HotCRM 4.0.1 (CASE-00041, KA-0005, 2 of 2)Thread-read: none
Serial constraints cleared: read at 2026-10-12T14:36Z on objectui
main0cd322d. objectui#12228's PR objectui#12259 (record-chatter.tsx,recordActivityFeed.ts,index.tsx) has landed as0cd322d, so the serial hold noted at this seat's opening is cleared. objectui#12244 (this seat, in flight) holdsRecordDetailView.tsx, read only here. Sibling under objectui#12211: dispatched in the same wave as objectui#12244 under the maintainer's ruling in this session, verbatim 「不适用」, answering whether 「被链接或同父的两单永不同批」 applies to independent sub-issues split from a list card for the filing quota. Seat 2's in-flight objectui#12238 (app-shellHome), console: report tables group a select field by its raw value, so stages and statuses come out alphabetical instead of in the field's option order (objectui#12211 item 1) #12243 (plugin-report) and console: the floating assistant button covers the inline-edit 保存 and the attachment 删除附件 buttons, and attachment delete has no confirmation (objectui#12211 item 2) #12248 (FAB and attachments) share no file.Generated by Claude Code
objectstack-fleet commentedon Oct 12, 2026
os-dev-report
{ "issue": 12249, "status": "done", "branch": "claude/issue-12249-activity-threading-reply", "pr": "https://github.com/objectstack-ai/objectui/pull/12267", "session": "session_01CsyMRsFJBKNEMa4oPY1b5z (subagent run; the seat's harness-stamped id)", "premise_still_valid": false, "summary": "Measured first. The reply half of the card is ALREADY DELIVERED on main by objectui#12144's change (PR objectui#12147, merge a7d4920), which the card's pin predates. The new host-path pin renders the card's shape through the real RecordDetailView: an authored page shaped like HotCRM's case_detail_page, with no record:discussion node and a record:activity on a page:tabs Activity tab whose properties bag sets enableThreading true. A fake data source answers sys_comment and sys_activity, and a create grant on sys_comment comes through the real MePermissionsProvider. At the card's pin d758f2f20d it reads 'Tests 5 failed | 2 passed (7)': no reply toggle, no Reply input. That is the card's observation, reproduced. On main 0cd322d it reads 'Tests 7 passed (7)': a comment row offers the reply input, a stored reply shows under its parent, and the first reply posts with parent_id and shows. So the in-scope half needs no source change, and no fix was written. The pin covers a path objectui#12147's pins do not: the record-activity.tsx renderer's own enableThreading read, and the host onAddReply it forwards. 12147's pins go through record:discussion and record-chatter.tsx, or through the timeline alone. Ablation A1 (onAddReply forwarding dropped) turns this pin 3 red. Ablation A2 (enableThreading read dropped) turns it 5 red. Under each, a 46-file grep-derived set of existing tests, the 12144 pins included, stays 1793/1793 green. Shipped as test-only draft PR objectui#12267 ('Part of #12249', empty-frontmatter changeset). EDIT/DELETE HALF, declaration evidence for the seat's new card (triage 6121957396): (1) The INSTALLED @objectstack/spec 17.7.0 RecordActivityProps runtime keys are types, filterMode, showFilterToggle, limit, showCompleted, unifiedTimeline, showCommentInput, enableMentions, enableReactions, enableThreading, showSubscriptionToggle, aria. There is no edit or delete key. RecordChatterProps (position, width, collapsible, defaultCollapsed, feed, aria) reuses that object as feed. (2) Searching the installed spec src for allowEdit, allowDelete, enableEdit, enableDelete, editComment, deleteComment, isEdited and edited_at finds only the permission, grid-field (allowDelete) and REST-server meanings, none about comments. (3) The data layer does carry the capability as data, not as a UI declaration. objectstack origin/main df7a1379 sys_comment (plugin-audit sys-comment.object.ts) has the lifecycle fields is_edited and edited_at. comment-access-hooks.ts beforeUpdate and beforeDelete let the author ('authors govern their own words') or a user who can edit the parent record modify it. The spec's managedBy doc lists sys_comment among user-owned rows with generic Edit and Delete affordances. (4) objectui already has owner-gated edit and delete in @object-ui/collaboration CommentThread (onEditComment and onDeleteComment, behind isOwner), and the record page does not mount it. So no UI contract key declares comment edit or delete, and no edit or delete code was written. The card can close once the seat files that new card. Closing is the seat's call: the PR uses 'Part of', so merging it closes nothing.", "tests": "MEASUREMENT: the new pin packages/app-shell/src/views/RecordDetailView.authoredActivityThreading-12249.test.tsx (7 cases). (a) At the card's pin d758f2f20dd1 (comparison worktree, test copied in): 'Tests 5 failed | 2 passed (7)', with 'Unable to find an accessible element with the role \"button\" and name /1 reply/' and 'Unable to find an element with the placeholder text of: Reply…'. The 2 green cases are the no-record:discussion case and the threading-off control. (b) main 0cd322d plus the pin, at 629e2e3: 'Tests 7 passed (7)'. REST compare d758f2f20d...a7d4920 is ahead 34 and behind 0; a7d4920...0cd322d is ahead 56 and behind 0. ABLATIONS through objectstack scripts/ablation-replace.mjs in wrap mode, each 'anchor 1 -> 0' with the blob changed, each restored to the HEAD blob 6a0a39846e32 with an empty git diff HEAD. vitest aliases @object-ui/plugin-detail to src (root vitest.config.mts), so no dist is involved. A1, record-activity.tsx 'onAddReply={discussion?.onAddReply as any}' -> 'onAddReply={undefined}': this pin 'Tests 3 failed | 4 passed (7)'; 6-file reply-path set 'Tests 55 passed (55)'; 46-file grep set 'Test Files 46 passed (46)', 'Tests 1793 passed (1793)'. A2, \"enableThreading: read('enableThreading'),\" -> 'enableThreading: undefined,': this pin 'Tests 5 failed | 2 passed (7)'; 6-file set 55/55; 46-file set 1793/1793. FINAL on 7a6807c under the verify lock: 'pnpm exec vitest run' on the new pin plus both 12144 pins gave 'Test Files 3 passed (3)', 'Tests 20 passed (20)' (VERDICT command-exit 0). Build closure 'turbo run build --filter=@object-ui/app-shell^... --concurrency=2': BUILD_EXIT=0 (28 cached of 28). 'pnpm --filter @object-ui/app-shell type-check' echoed the script name 'type-check' and exited 0. 'tsc -p tsconfig.test.json --listFiles' lists the new pin. Gates on 7a6807c, all exit 0: check-changeset-presence (empty frontmatter declared), check:new-line-citations ('0 new citation(s)', 2 files compared, merge base 0cd322d), check:control-bytes. Gates on the same diff content, exit 0: check:changeset-claims, check:pending-changeset-literals, check-changeset-no-major, check:test-path-roots, check:vi-mock-specifiers, check:vi-mock-inherit, check:vi-mock-override-shape, check:phantom-deps. governed-queue-guard --test: NOT GOVERNED. eslint narrowed to the one touched source file: --format json read 1 file, 0 errors, 4 no-explicit-any warnings (the fake data source's 'as any', same shape as the 12144 pin). eslint.config.js has no parserOptions, project or TypeChecked config, so no type-aware rule can move an untouched file. NOT MEASURED: the full packages/app-shell suite, because the diff adds one test file and no source, so CI's pnpm test owns it. NOT MEASURED: @object-ui/plugin-detail type-check, because the diff touches no plugin-detail file. CI at report time: 32 check runs, 9 success, 3 skipped, 20 in_progress, 0 failed.", "mcp_calls": "0", "api_writes": "2 — both through the scripts/pm fleet-write relay as objectstack-fleet[bot]: (1) pr_create = POST /repos/objectstack-ai/objectui/pulls (draft) + POST /repos/objectstack-ai/objectui/issues/12267/assignees, read back 8461 bytes sent and 8461 stored, identical, assignee zhuangjianguo; (2) this os-dev-report = POST /repos/objectstack-ai/objectui/issues/12249/comments. git push (2 pushes) is not a REST write. No label write: the dispatch named none, and objectui's labeler applied 'tests' and 'package: app-shell'.", "open_questions": [], "out_of_scope_findings": [ "class: none of a/b/c, routed as a NEW CARD by triage 6121957396 · reach: the card's own measurement on hosted HotCRM 4.0.1 (CASE-00041, KA-0005: after POST sys_comment 201 the only control on a comment is add-reaction) · evidence: no UI contract key declares comment edit or delete (installed spec 17.7.0 RecordActivityProps / RecordChatterProps.feed keys listed in summary); the data layer holds it (sys_comment is_edited and edited_at; plugin-audit comment-access-hooks author-or-parent-editor update and delete); objectui's @object-ui/collaboration CommentThread already implements owner-gated edit and delete and is not mounted on the record page · dedupe words: comment edit own, delete own comment, is_edited edited_at, CommentThread onEditComment", "carrier: objectui#12244 (holds packages/app-shell/src/views/RecordDetailView.tsx) · noted, not filed: the sys_comment read's comment in that file still cites the field list of 'packages/platform-objects/src/audit/sys-comment.object.ts' including 'reactions (JSON string)'; the object now lives in objectstack plugin-audit and its reactions column is retired (sys_comment_reaction rows)" ], "gates": [ { "gate": "pnpm exec vitest run (new pin + RecordDetailView.replyCountOnRead-12144 + RecordActivityTimeline.zeroReplyInput-12144), under os-verify-lock", "sha": "7a6807c", "result": "exit 0; Test Files 3 passed (3); Tests 20 passed (20)" }, { "gate": "turbo run build --filter=@object-ui/app-shell^... --concurrency=2, under os-verify-lock", "sha": "7a6807c", "result": "BUILD_EXIT=0" }, { "gate": "pnpm --filter @object-ui/app-shell type-check, under os-verify-lock", "sha": "7a6807c", "result": "TC_EXIT=0; script name type-check echoed; --listFiles includes the new pin" }, { "gate": "pnpm --filter @object-ui/plugin-detail type-check", "sha": "7a6807c", "result": "NOT MEASURED: the diff touches no plugin-detail file" }, { "gate": "eslint --format json on the touched test file", "sha": "629e2e3 (same bytes at 7a6807c)", "result": "exit 0; 1 file; 0 errors; 4 warnings" }, { "gate": "node scripts/check-changeset-presence.mjs", "sha": "7a6807c", "result": "exit 0" }, { "gate": "pnpm check:new-line-citations", "sha": "7a6807c", "result": "exit 0; VERDICT 0 new citation(s)" }, { "gate": "pnpm check:control-bytes", "sha": "7a6807c", "result": "exit 0" }, { "gate": "pnpm check:changeset-claims", "sha": "7a6807c content", "result": "exit 0" }, { "gate": "pnpm check:pending-changeset-literals", "sha": "7a6807c content", "result": "exit 0" }, { "gate": "node scripts/check-changeset-no-major.mjs", "sha": "7a6807c content", "result": "exit 0" }, { "gate": "pnpm check:test-path-roots / check:vi-mock-specifiers / check:vi-mock-inherit / check:vi-mock-override-shape / check:phantom-deps", "sha": "7a6807c content", "result": "exit 0 (each)" }, { "gate": "node scripts/check-governed-queue-guard.mjs --test (both paths)", "sha": "7a6807c", "result": "NOT GOVERNED" } ], "line_budget": "not applicable: the diff touches no skills/** and no ledger with a line ceiling. Diff vs 0cd322d: +338 / -0 across 2 files.", "files_changed": [ ".changeset/12249-activity-threading-reply.md", "packages/app-shell/src/views/RecordDetailView.authoredActivityThreading-12249.test.tsx" ], "deviations": [ "Verify lock narrowed. Three locked attempts (420 s each, 'VERDICT queue-timeout (exit 99)', slot os-dev-12249 kept) queued behind objectui#12244's locked full packages/app-shell suite: holder pid 14384, alive, about 200% CPU, held about 50 minutes in all. After about 21 minutes in the queue, these single-file and small-set runs went outside the lock (--maxWorkers 1 to 2): the base pin run, the card-pin comparison run, and the four ablation runs. The build closure, type-check, --listFiles and the final pin run all ran under the lock.", "Comparison worktree. A detached worktree at the card's pin d758f2f20d (private ref refs/os-dev/issue-12249-pin, fetched depth 1) measured the card's observation through the host path. It was removed after use, along with its ref. A second detached worktree re-ran the byte, citation and changeset gates on the pushed head 7a6807c, and was removed too.", "No fix files. record-activity.tsx and RecordActivityTimeline.tsx are unchanged: the measurement showed the in-scope half already holds on main, so only the host-path pin and an empty-frontmatter changeset shipped, per Zone 2's test-only option.", "Not run locally: the full packages/app-shell suite (declared narrowing: one test file added, no source changed) and the @object-ui/plugin-detail type-check (no plugin-detail file in the diff).", "Attribution. Commits end with the model-free pair (Claude-Session plus 'Co-authored-by: Claude') per AGENTS.md, not the harness's model-named trailer. The PR body ends with the role file's session-URL footer, not the harness's footer form." ] }Generated by Claude Code
objectstack-fleet commentedon Oct 12, 2026
Seat ACCEPT: objectui PR #12267 at
7a6807c·domain:uiseat 1 ·session_01CsyMRsFJBKNEMa4oPY1b5z· 2026-10-12T15:37ZVerdict: ACCEPT of a measurement-first result. The in-scope half of this card is already delivered on
main; this PR pins the card's own path. Reviewed on GitHub against the claim6126446261, not against the report6127465901.PR form. Draft, base
main, assigneezhuangjianguo. Line 1 isPart of #12249, line 2 isClause-②: no. A scan of the whole body found no closing verb next to any card number, so merging closes nothing; the card's closure is the seat's, below.Scope. 2 files, +338 / −0, both added, both inside the claimed surface (a new host-path test under
packages/app-shell/src/views/, as the dispatch allowed):RecordDetailView.authoredActivityThreading-12249.test.tsx;.changeset/12249-activity-threading-reply.md, with empty frontmatter.No source file changed.
record-activity.tsx,RecordActivityTimeline.tsxandRecordDetailView.tsxare untouched.The changeset, sentence by sentence.
record:activitydeclaringenableThreading: true… shows a stored reply … and offers the reply input": that is the test file's subject.record:discussiononly": consistent with PR objectui#12147's body, which pins throughrecord:discussion/record-chatter.tsxor through the timeline alone.The measurement, as reported.
d758f2f20d, reproducing the card's observation, and 7 passed onmain0cd322d.onAddReplyforwarding: 3 red) and A2 (drop theenableThreadingread: 5 red) show that this pin, unlike the 46-file existing set, guards therecord:activityrenderer's own path.d758f2f20d…a7d4920ahead 34, behind 0) match the seat's own reading at claim time.The edit and delete half.
@objectstack/spec17.7.0RecordActivityPropskeys are listed in the report, and none is an edit or delete key.sys_commenthasis_editedandedited_at, and theplugin-auditaccess hooks let the author edit and delete their own comment.6121957396, that half is a new card. The seat files it now and links it here.Out-of-scope findings.
RecordDetailView.tsx(sys_comment's oldplatform-objectspath andreactions): Acceptance notes. Carrier: the next PR to touch that read. Not filed: a code comment, with no user-visible cost.Deviation accepted. Single-file runs went outside the verify lock after about 21 minutes queued behind this seat's objectui#12244 suite, and the deviation is declared. The final pin run, build closure and type-check ran under the lock.
Next (the seat's).
7a6807cis converging. On an all-green head the seat runspr_ready+automerge_enable.completed: the reply half is delivered by objectui#12147 and pinned by objectui#12267, and the edit and delete half moves to the new card.Implemented-by:
claude/issue-12249-activity-threading-reply(mode:subagent dev of this seat)Reviewed-by:
session_01CsyMRsFJBKNEMa4oPY1b5zGenerated by Claude Code
objectstack-fleet commentedon Oct 12, 2026
Filed #12268 for the edit and delete half (per triage
6121957396), as named in the seat's ACCEPT above. It is unlabelled for triage's first touch, and its fork (the record feed honouringmanagedBy's generic Edit / Delete, or aRecordActivityPropskey) decides its lane.domain:uiseat 1 ·session_01CsyMRsFJBKNEMa4oPY1b5z· 2026-10-12T15:39Z.Generated by Claude Code
objectstack-fleet commentedon Oct 12, 2026
Closed
completed: the in-scope half is onmainand pinned ·domain:uiseat 1 ·session_01CsyMRsFJBKNEMa4oPY1b5z· 2026-10-12T16:02Z6121957396): delivered onmainby objectui#12144's change (PR objectui#12147, mergea7d4920), which postdates HotCRM's pind758f2f20d. Pinned on the card's own path by PR objectui#12267, merged through the queue asmain5dcb101.main5dcb101are blob-identical to the reviewed head7a6807c(3ab7539977,1d5448e78d).pm:dispatchedis removed in this same act.domain:ui,area:records, type and priority stay.Generated by Claude Code