You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository was archived by the owner on Nov 4, 2026. It is now read-only.
Repository navigation
This repository was archived by the owner on Nov 4, 2026. It is now read-only.
plugin-map 的 Features 清单与 ObjectMap.tsx 头注释仍宣称 api data provider 可用(实为未实现分支) #5163
Claim: PM loop round 1
Session: session_01AMHbqfPiETJHA95r6WzZWS
Branch: claude/issue-5163-map-api-provider-claim
Worktree: objectui-issue-5163
Domain: repo:objectui (execution seat; this repo has no domain:* lanes)
File surface: exactly two lines — content/docs/plugins/plugin-map.mdx:24 and packages/plugin-map/src/ObjectMap.tsx:20 — plus a changeset (stop on breach; explain in the report)
Container & model: S mechanical, mode:subagent, model: sonnet — dispatch-gates.mjs --tier reports no path-derived mandate; this sits on the floor deliberately: two one-line text edits whose correctness the doc gates decide mechanically, with no design judgment left in the card.
Clause-②: no — prose and a source comment only; no code path, no contract, no public surface.
Serial constraints cleared: all three blockers are now on main. The gating comment (comment-5325556976) required PR #5156 and PR #5162 both MERGED; both are (578e02516, 7956b9c25). Since that comment was written, PR #5175 (#5017) also landed on ObjectMap.tsx (072085deb, main HEAD) — I am recording it because it was not in the original gating list and it is the reason this card was still not dispatchable an hour ago. ObjectMap.tsx now carries no in-flight claim.
Premise re-measured against origin/main @ 072085de — both targets still exist, both still at the line numbers the card names:
content/docs/plugins/plugin-map.mdx:24
- **ObjectQL Integration**: Works seamlessly with object/api/value data providers
packages/plugin-map/src/ObjectMap.tsx:20
* - Works with object/api/value data providers
and the branch that makes them false is still there — but it has moved, so take line numbers from the tree, not from the card:
packages/plugin-map/src/ObjectMap.tsx:584 console.warn('API provider not yet implemented for ObjectMap');
The card says :475. PR #5156 and PR #5175 pushed it to :584. git grep -n "endpoint" -- packages/plugin-map/src is still a genuine zero (reverse-checked: neighbouring strings in the same files do match, so the zero is real and not a broken pattern).
Scope — this is where the card can go wrong, so read it before you edit. The string object/api/value appears at seven sites in this repo. Yours is two of them. I grepped the rest during the premise check and they are not all the same defect:
plugin-gantt (plugin-gantt.mdx:24, ObjectGantt.tsx:22) — the identical sentence, and there it is TRUE.provider: 'api' really is implemented for gantt (ObjectGantt.tsx:442-445, :1155).
So: ⛔ do not grep-and-replace object/api/value repo-wide. It would silently make the gantt docs wrong. Touch the two plugin-map lines and nothing else; if you think a sibling site belongs in this PR, stop and say so instead.
Both edits are object/api/value → object/value. The card explicitly does not propose implementing provider: 'api' for map — that is capability expansion and is out of scope.
{
"issue": 5163,
"status": "done",
"branch": "claude/issue-5163-map-api-provider-claim",
"pr": "https://github.com/objectstack-ai/objectui/pull/5181",
"premise_still_valid": true,
"summary": "Corrected the two remaining false object/api/value claims for plugin-map that PR #5162 (issue #5019) left out of its scope: the Features list line in content/docs/plugins/plugin-map.mdx:24 and the file-header JSDoc in packages/plugin-map/src/ObjectMap.tsx:20, both changed to object/value. Left plugin-calendar (same defect, filed separately as #5180) and plugin-gantt (same sentence but TRUE there, provider: 'api' is implemented) untouched, as instructed. Diff is exactly the two lines plus a changeset.",
"tests": "All commands run at head 09c0a385c (matches PR branch). node scripts/check-control-bytes.mjs -> pass (4591 files). node scripts/check-doc-links.mjs -> pass (13 scan roots). node scripts/check-doc-component-types.mjs -> pass. node scripts/check-doc-snippet-types.mjs --build-filter -> plugin-map absent from derived filter, so Doc Snippet Type Check does not exercise this diff (confirmed, not assumed). pnpm turbo run build --filter='@object-ui/site' --concurrency=2 (Build Docs) -> 29/29 tasks succeeded incl. edited plugin-map.mdx page, 3m30s. pnpm --filter '@object-ui/plugin-map^...' build then pnpm --filter '@object-ui/plugin-map' build -> both exit 0 (dependency closure built before judging anything). pnpm --filter '@object-ui/plugin-map' type-check (tsc --noEmit && tsc -p tsconfig.test.json) -> pass. pnpm exec vitest run packages/plugin-map/ --maxWorkers=2 (repo-root invocation per AGENTS.md) -> 9 test files / 61 tests passed. node scripts/check-changeset-presence.mjs -> pass, 1 changeset declared for 1 released source file changed. node scripts/check-changeset-no-major.mjs -> pass. All heavy commands ran serialized under /tmp/os-heavy-verify.lock.",
"open_questions": [],
"out_of_scope_findings": []
}
Checklist verdict, checked against GitHub rather than the report's self-description.
Scope — the one thing that could have gone wrong here, and didn't.get_files returns exactly three: the changeset, content/docs/plugins/plugin-map.mdx (+1/−1), packages/plugin-map/src/ObjectMap.tsx (+1/−1). Both edits are literally object/api/value → object/value. No plugin-gantt, no plugin-calendar, no repo-wide sweep of the string. That was the live hazard on this card — five sibling sites carry the same sentence and one family (gantt) is correctly claiming an implemented api provider, so a grep-and-replace would have shipped a fresh docs bug while looking thorough. The dev checked all seven sites and touched two.
Spot-checks (my own readings, not the report's):
Base is 072085deb = current main head. No stale-baseline risk.
First line Fixes #5163 — correct: this card is fully closed by these two lines, nothing residual, so a closing keyword is right rather than Part of.
The dev independently caught that the card's :475 for the not yet implemented branch is stale and is now :584 — matching what I measured at dispatch. Line numbers were taken from the tree, as instructed.
endpoint / method zero read points in packages/plugin-map/src — I confirmed this myself at dispatch with a reverse-check against neighbouring strings, so the zero is real rather than a broken pattern.
Diff is two lines with no reflow or re-wrap, which is what keeps this card reviewable at a glance.
Honest reading recorded rather than glossed: the dev reports check-doc-snippet-types.mjs --build-filter does not exercise this diff (plugin-map is absent from the derived filter) and says so explicitly instead of listing it as a passing gate. That is the right way to report a gate that does not apply — a green that proves nothing is not evidence.
Landing path: path surface is .changeset/** + content/docs/** + packages/plugin-map/src/** — no governed face touched, so this is queue-eligible rather than human-merge-only.
CI not yet converged at review time and that is an honest state, not a pass: 12 checks terminal (10 success, 2 skipped) with 8 still in_progress (Lint, Type Check, four test shards, Build Docs, Doc Snippet Type Check). Convergence is the reviewer's side of the contract, not the dev's — I flip ready and enqueue only once every required job reaches a terminal success, and I read the named jobs directly rather than an aggregate. If anything lands red I reopen this as REWORK on the same claim.
Scope note for the record: the sibling defect this card deliberately excluded is tracked as #5180 (plugin-calendar, three sites, same stub shape) — filed unassigned and ungraded, awaiting the triage seat.
Activity
yinlianghui commentedon Aug 18, 2026
PM 分诊(objectui 分片 PM,session
session_01GTRjn8xBqp75dk7kFupVRt):晋级pm:queue,S 级两行改动(object/api/value→object/value×2)。串行约束(派发条件):两处目标一在
plugin-map.mdx、一在ObjectMap.tsx:20头注释 —— 后者在 PR #5156(#5018 实施,+127/−76 重写该文件)的活动面上,前者所在页同时有 PR #5162(#5019)在途。两个 PR 都 MERGED 后再派,基线取含两者的 main 实 sha,届时按落地后的行号重新定位。不派前不占席。Generated by Claude Code
os-steve commentedon Aug 18, 2026
Claim: PM loop round 1
Session:
session_01AMHbqfPiETJHA95r6WzZWSBranch:
claude/issue-5163-map-api-provider-claimWorktree:
objectui-issue-5163Domain:
repo:objectui(execution seat; this repo has nodomain:*lanes)File surface: exactly two lines —
content/docs/plugins/plugin-map.mdx:24andpackages/plugin-map/src/ObjectMap.tsx:20— plus a changeset (stop on breach; explain in the report)Container & model:
Smechanical,mode:subagent,model: sonnet—dispatch-gates.mjs --tierreports no path-derived mandate; this sits on the floor deliberately: two one-line text edits whose correctness the doc gates decide mechanically, with no design judgment left in the card.Clause-②: no — prose and a source comment only; no code path, no contract, no public surface.
Serial constraints cleared: all three blockers are now on main. The gating comment (comment-5325556976) required PR #5156 and PR #5162 both MERGED; both are (
578e02516,7956b9c25). Since that comment was written, PR #5175 (#5017) also landed onObjectMap.tsx(072085deb, main HEAD) — I am recording it because it was not in the original gating list and it is the reason this card was still not dispatchable an hour ago.ObjectMap.tsxnow carries no in-flight claim.Premise re-measured against
origin/main@072085de— both targets still exist, both still at the line numbers the card names:and the branch that makes them false is still there — but it has moved, so take line numbers from the tree, not from the card:
The card says
:475. PR #5156 and PR #5175 pushed it to :584.git grep -n "endpoint" -- packages/plugin-map/srcis still a genuine zero (reverse-checked: neighbouring strings in the same files do match, so the zero is real and not a broken pattern).Scope — this is where the card can go wrong, so read it before you edit. The string
object/api/valueappears at seven sites in this repo. Yours is two of them. I grepped the rest during the premise check and they are not all the same defect:plugin-calendar(plugin-calendar.mdx:178,ObjectCalendar.tsx:22,index.md:168) — same false claim, same stub (ObjectCalendar.tsx:295). Out of your surface; filed separately as finding(plugin-calendar): the same false "api data provider works" claim #5163 records for plugin-map exists at three calendar sites — and the gantt copy of the sentence is TRUE, so this is not a blanket sweep #5180.plugin-gantt(plugin-gantt.mdx:24,ObjectGantt.tsx:22) — the identical sentence, and there it is TRUE.provider: 'api'really is implemented for gantt (ObjectGantt.tsx:442-445,:1155).So: ⛔ do not grep-and-replace
object/api/valuerepo-wide. It would silently make the gantt docs wrong. Touch the two plugin-map lines and nothing else; if you think a sibling site belongs in this PR, stop and say so instead.Both edits are
object/api/value→object/value. The card explicitly does not propose implementingprovider: 'api'for map — that is capability expansion and is out of scope.Generated by Claude Code
fix(plugin-map): correct residual "object/api/value" claim to "object…
os-steve commentedon Aug 18, 2026
{ "issue": 5163, "status": "done", "branch": "claude/issue-5163-map-api-provider-claim", "pr": "https://github.com/objectstack-ai/objectui/pull/5181", "premise_still_valid": true, "summary": "Corrected the two remaining false object/api/value claims for plugin-map that PR #5162 (issue #5019) left out of its scope: the Features list line in content/docs/plugins/plugin-map.mdx:24 and the file-header JSDoc in packages/plugin-map/src/ObjectMap.tsx:20, both changed to object/value. Left plugin-calendar (same defect, filed separately as #5180) and plugin-gantt (same sentence but TRUE there, provider: 'api' is implemented) untouched, as instructed. Diff is exactly the two lines plus a changeset.", "tests": "All commands run at head 09c0a385c (matches PR branch). node scripts/check-control-bytes.mjs -> pass (4591 files). node scripts/check-doc-links.mjs -> pass (13 scan roots). node scripts/check-doc-component-types.mjs -> pass. node scripts/check-doc-snippet-types.mjs --build-filter -> plugin-map absent from derived filter, so Doc Snippet Type Check does not exercise this diff (confirmed, not assumed). pnpm turbo run build --filter='@object-ui/site' --concurrency=2 (Build Docs) -> 29/29 tasks succeeded incl. edited plugin-map.mdx page, 3m30s. pnpm --filter '@object-ui/plugin-map^...' build then pnpm --filter '@object-ui/plugin-map' build -> both exit 0 (dependency closure built before judging anything). pnpm --filter '@object-ui/plugin-map' type-check (tsc --noEmit && tsc -p tsconfig.test.json) -> pass. pnpm exec vitest run packages/plugin-map/ --maxWorkers=2 (repo-root invocation per AGENTS.md) -> 9 test files / 61 tests passed. node scripts/check-changeset-presence.mjs -> pass, 1 changeset declared for 1 released source file changed. node scripts/check-changeset-no-major.mjs -> pass. All heavy commands ran serialized under /tmp/os-heavy-verify.lock.", "open_questions": [], "out_of_scope_findings": [] }Generated by Claude Code
os-steve commentedon Aug 18, 2026
ACCEPT — PR #5181 (
repo:objectuiseat PM, sessionsession_01AMHbqfPiETJHA95r6WzZWS, reviewer of record).Checklist verdict, checked against GitHub rather than the report's self-description.
Scope — the one thing that could have gone wrong here, and didn't.
get_filesreturns exactly three: the changeset,content/docs/plugins/plugin-map.mdx(+1/−1),packages/plugin-map/src/ObjectMap.tsx(+1/−1). Both edits are literallyobject/api/value→object/value. No plugin-gantt, no plugin-calendar, no repo-wide sweep of the string. That was the live hazard on this card — five sibling sites carry the same sentence and one family (gantt) is correctly claiming an implementedapiprovider, so a grep-and-replace would have shipped a fresh docs bug while looking thorough. The dev checked all seven sites and touched two.Spot-checks (my own readings, not the report's):
072085deb= currentmainhead. No stale-baseline risk.Fixes #5163— correct: this card is fully closed by these two lines, nothing residual, so a closing keyword is right rather thanPart of.:475for thenot yet implementedbranch is stale and is now:584— matching what I measured at dispatch. Line numbers were taken from the tree, as instructed.endpoint/methodzero read points inpackages/plugin-map/src— I confirmed this myself at dispatch with a reverse-check against neighbouring strings, so the zero is real rather than a broken pattern.Honest reading recorded rather than glossed: the dev reports
check-doc-snippet-types.mjs --build-filterdoes not exercise this diff (plugin-map is absent from the derived filter) and says so explicitly instead of listing it as a passing gate. That is the right way to report a gate that does not apply — a green that proves nothing is not evidence.Deviations: none. File surface as claimed, changeset present (
@object-ui/plugin-mappatch), no-major gate green.Landing path: path surface is
.changeset/**+content/docs/**+packages/plugin-map/src/**— no governed face touched, so this is queue-eligible rather than human-merge-only.CI not yet converged at review time and that is an honest state, not a pass: 12 checks terminal (10 success, 2 skipped) with 8 still
in_progress(Lint, Type Check, four test shards, Build Docs, Doc Snippet Type Check). Convergence is the reviewer's side of the contract, not the dev's — I flip ready and enqueue only once every required job reaches a terminalsuccess, and I read the named jobs directly rather than an aggregate. If anything lands red I reopen this as REWORK on the same claim.Scope note for the record: the sibling defect this card deliberately excluded is tracked as #5180 (plugin-calendar, three sites, same stub shape) — filed unassigned and ungraded, awaiting the triage seat.
Generated by Claude Code
fix(plugin-map): correct residual "object/api/value" claim to "object…