Repository navigation
A sales rep cannot save a new lead without picking themselves as owner: the Create Lead form marks owner_id required, which blocks the insert-time stamp that already makes the creator the owner #2097
Description
Filing gate ①: a user-visible defect, reproduced in a browser.
reach: the console Leads → New, as na.rep.
Filed by the repo:hotcrm seat (session_01BCDqr1DqnQjEh5jA9gGkUG) from the v18 test pass #2093 (finding 08), under the maintainer's 「测试遇到问题就报issue」. ⛔ Not a claim.
Reader: this seat's queue.
Dedupe: the hotcrm index on owner required / lead owner / owner prefill: 0. #2060's own dedupe line records Lead Owner is required → 0 hits.
What happens (main 2f3d898d, @objectstack/* 18.0.0-next.0)
As na.rep: open Leads → New, fill First / Last / Company / Email, and click Create.
- Result: the inline error "Lead Owner is required" appears, nothing is POSTed, and the rep has to search for and pick themselves.
- The Opportunity, Account and Contact New forms also leave Owner blank ("Select…"), but none of them requires it.
Screenshot: 08-create-lead-form-owner-required-not-prefilled.png (on #2093's branch).
Cause (seat-verified in source)
src/sales/views/lead.view.ts~293–296 puts{ field: 'owner_id', required: true }in the create form.src/sales/objects/lead.hook.ts~41–43 states that "UI-created leads already carryowner_id= creator (the middleware's insert-time stamp)". The form'srequiredstops the save before that stamp can apply.
Fix direction (measure first)
- Drop
required: truefrom the create form'sowner_id, so the insert-time stamp makes the creator the owner, as on the other three objects. - Re-measure that a rep's new lead saves owned by the rep, and that a manager can still pick another owner.
- Check
git log -Sfor whyrequiredwas added. If a ruling put it there, stop and report. - That the console never prefills
owner_idwith the current user is the platform half: noted, not part of this card.
Generated by Claude Code
Activity
objectstack-fleet commentedon Oct 14, 2026
Claim: PM loop round R85
Session:
session_01P8mZuzuyZAHSrgQKmv6ASvAccount:
sunhaolin(the seat's linked user asGET /useranswers it; always the card's assignee)Branch:
claude/issue-2097-lead-owner-not-requiredWorktree:
hotcrm-issue-2097Domain:
repo:hotcrm(single-lane repo, nodomain:*taxonomy)Seat:
repo:hotcrm#1File surface:
src/sales/views/lead.view.ts(the create form'sowner_identry only, ~L293–296, and its comment);content/docs/**/leads*.mdxonly if a sentence says a new lead needs an owner picked; tests undertest/; one.changeset/2097-*.md. A producer measured in the platform (the insert-time owner stamp not applying, or the console's form behaviour) goes upstream and this card WAITS (stop on breach; explain in the report)Container & model:
S(one form entry; the judgment is thegit log -Sprovenance check and the re-measure),mode:subagent,model: default(dispatch-gates --tier --repo objectstack-ai/hotcrm src/sales/views/lead.view.ts: "no path-derived mandate: the surface hits none of the 3 declared glob(s)"; the tier stays the PM's per-card call)Clause-②: no
Responsibility:
the app: lead.view.ts's create form sets owner_id required: true, which refuses the save before the insert-time stamp applies | the platform path that already covers it: the insert-time owner stamp that lead.hook.ts ~41–43 records | every rep who creates a lead in the console, todayThread-read: none
Serial constraints cleared:
src/sales/views/lead.view.tsis shared with #2094, claimed in this round. Fold-or-serial: not folded (gate ① fails: different defect shapes). Parallel by region: this card owns the create form'sowner_identry (~L293–296), #2094 owns the listcolumnsblock (~L141–200); the later lander mergesmain. Last change on the file:81f8644(#2050), not today. In flight in this lane: #2094 and #2095 (this round). Open PRs: dependabot #2063 and #2064 only.Generated by Claude Code
objectstack-fleet commentedon Oct 14, 2026
os-dev-report
{
"issue": 2097,
"status": "done",
"branch": "claude/issue-2097-lead-owner-not-required",
"pr": "#2099",
"session": "session_01P8mZuzuyZAHSrgQKmv6ASv (this run is a subagent; the session id is the parent seat's, as the claim and the commit trailers carry it)",
"premise_still_valid": true,
"summary": "Removed
required: truefrom theowner_identry of the lead create form (LeadViews.formin src/sales/views/lead.view.ts). The field stays on the form as a bareowner_idwith a comment saying why, which matches the Opportunity, Account and Contact New forms. Added one test in test/ownership-model.test.ts (form has owner_id, not required) and a patch changeset. H1 held: crm_lead.owner_id is Field.lookup sys_user with system: true and readonly: false, not required, and no validations[] or requiredWhen touches it; once the flag was gone the same rep save went through. H2 held, measured in a way that tells the two mechanisms apart: the sales_rep pool is [na.rep, eu.rep] with 1 and 0 open leads, so round-robin would have picked eu.rep, yet na.rep's second ownerless lead was saved owned by na.rep. sales.manager, who is not in the pool, also got their own lead. H4: no ruling.requiredcame from the 344c5f5 framework/app-crm baseline on the oldownerfield, and 9f748ab (#548 via #677) renamed it to owner_id mechanically, so the stop line did not trigger. H3 FAILED, but for a platform reason that exists on base too: when the manager picks Nina Reyes, the chip shows the pick, yet the POST body has no owner_id and the lead is saved under the manager. This is identical on base 2f3d898 (owner required) and on this branch. The producer is objectui plugin-form: the create branch strips owner_id, and only an edit adds back a changed owner (objectui#12205). Reported in out_of_scope_findings, not worked around. I usedFixes #2097because the card's own defect (a rep blocked from saving) is fully fixed in this repo. If the seat rules that the card waits on the upstream manager-pick finding, the PR body's first line should change toPart of #2097(seat edit; dev writes the body once).","tests": "All on HEAD ca836d0. pnpm validate: "✓ Validation passed (2773ms)", exit 0, lock VERDICT command-exit 0. pnpm typecheck: exit 0; tsc --listFiles includes test/ownership-model.test.ts and src/sales/views/lead.view.ts. pnpm lint: "18 suggestion(s)", exit 0, lock VERDICT command-exit 0. pnpm lint:i18n-gate: "✓ i18n lint gate: 0
i18n/missing-*issues", exit 0. pnpm hygiene: "✓ source hygiene clean", exit 0. pnpm hygiene:tokens: "✓ source token ratchet clean", exit 0. pnpm build: "✓ Build complete (2937ms)", exit 0, lock VERDICT command-exit 0. pnpm test: "Test Files 179 passed (179)", "Tests 3660 passed | 1 skipped (3661)", TEST_EXIT=0 captured right after pnpm test. The lock printedVERDICT batch-last-exit 0because the wrapped command waspnpm test ...; echo TEST_EXIT=$?, so TEST_EXIT is the verdict. Targeted run: vitest run test/ownership-model.test.ts gave "Tests 16 passed (16)", lock VERDICT command-exit 0 after a 7m52s queue wait. Reverse check (one-time): objectstack scripts/ablation-replace.mjs replaced the bare owner_id entry with{ field: 'owner_id', required: true }; the anchor hit x1, then x0, and the blob changed 049bedbe -> 251dc02d. Prediction stated before running: red. Observed: "Tests 1 failed | 15 passed (16)", "AssertionError: the New Lead form demands an owner again: expected true to be false", lock VERDICT command-exit 1. Restore: "ok restored: blob == HEAD (049bedbe81f3) and git diff HEAD is empty". The test imports src/ directly, so dist/ plays no part. Console measurements on port 4102 with Playwright DOM and network probes plus REST read-back, 18.0.0-next.0, demo:staff exit 0. BASE as na.rep: label "Lead Owner*", inline error "Lead Owner is required", 0 writes. HEAD as na.rep, lead 1: POST 201, no owner_id key in the body, read-back owner_id = na.rep (yH7k79Sc...). HEAD as na.rep, lead 2: owner = na.rep (round-robin would have chosen eu.rep). HEAD as sales.manager, no pick: owner = manager (fHtM0IOT...). HEAD as sales.manager picking Nina Reyes: chip "Remove Nina Reyes" shown, POST body without owner_id, owner = manager. BASE (rebuilt with the required form; artifact entry {"field":"owner_id","required":true}) as sales.manager picking Nina Reyes: same result. NOT MEASURED: Playwright e2e (CI's). CI at report time: Vercel Preview Comments and Label Pull Request success; Quality Checks, Build and Test (22.x), Check Changeset, Playwright, link-check and Analyze Code in_progress.","mcp_calls": "0 — no MCP GitHub tool was called",
"api_writes": "2 — (1) pr_create through the fleet relay: POST /repos/objectstack-ai/objectstack/dispatches (HTTP 204), relay run 38534739208 success, which created #2099 (draft, assignees [sunhaolin]); read-back 5484 bytes sent = 5484 stored. (2) this os-dev-report comment via scripts/pm/post-stamped.mjs --comment=2097. git push (4 pushes on the branch) is not REST and is not counted. Zero label writes.",
"gates": {
"head": "ca836d0d",
"pnpm validate": "exit 0 · ✓ Validation passed",
"pnpm typecheck": "exit 0",
"pnpm lint": "exit 0 · 18 suggestion(s)",
"pnpm lint:i18n-gate": "exit 0 · ✓ i18n lint gate: 0 i18n/missing-* issues",
"pnpm hygiene": "exit 0 · ✓ source hygiene clean",
"pnpm hygiene:tokens": "exit 0 · ✓ source token ratchet clean",
"pnpm build": "exit 0 · ✓ Build complete",
"pnpm test": "exit 0 · Test Files 179 passed (179) · Tests 3660 passed | 1 skipped (3661)",
"playwright e2e": "NOT MEASURED (CI)"
},
"line_budget": "net diff origin/main...HEAD: 3 files, +26 -4. src/sales/views/lead.view.ts +5 -4 (all comment or form-entry lines; hygiene:tokens unchanged, since comments are stripped and the entry got shorter). test/ownership-model.test.ts +16. .changeset/2097-lead-owner-not-required.md +5. No skills/** touched.",
"files_changed": [
"src/sales/views/lead.view.ts",
"test/ownership-model.test.ts",
".changeset/2097-lead-owner-not-required.md"
],
"deviations": [
"Commit 6b5c812 picked up
objectstack.config.bundled_4tw0mek20uf.mjs(35749 lines, a temporary compile file from the runningobjectstack devwatcher) because I usedgit add -A. Commit b5237b6 removes it, and the net PR diff is 3 files. It was scanned before removal: no absolute paths and no secret-shaped strings. History was not rewritten (no force-push). Later commits stage explicit paths.","Commit 6b5c812 carries the harness attribution pair, with a
Co-Authored-Byline that names the model. It was written following the harness reminder instead of the model-free pair, and was not rewritten. Commits b5237b6 and ca836d0 use the model-free pair (Claude-Session+Co-authored-by: Claude).","The console was booted with
pnpm exec objectstack dev -p 4102(its watcher compiles and rebuilds on change) without a separatepnpm buildfirst. The before and after states were measured on the same server across a watcher rebuild, and the artifact entry was read from dist/objectstack.json each time. The dev server process group (pgid 3636, started by me) was stopped with kill -TERM -3636, and port 4102 was confirmed free.","Ran
git fetch --unshallow origin mainin the shared /home/user/hotcrm .git so H4 could readgit log -Sand blame. The shared repo is now a full clone; this only adds history."],
"open_questions": [],
"out_of_scope_findings": [
"class: a · reach: console Leads → New as sales.manager@objectos.ai (sales_manager holds allowTransfer + modifyAllRecords on crm_lead). In Lead Owner, pick "Nina Reyes": the chip "Remove Nina Reyes" renders and no error is shown. Click Create: POST /api/v1/data/crm_lead returns 201 and its body carries no owner_id key, and the saved row's owner_id is the manager, not Nina Reyes. The picked owner is silently discarded. Identical on hotcrm base 2f3d898 (owner_id required) and on PR #2099, with @objectstack/console 18.0.0-next.0. · producer (read in objectui main b9814b3): packages/plugin-form/src/ModalForm.tsx create branch →
omitServerResolvedDefaults(sanitizeFormData(...)). sanitize.ts lists owner_id in SERVER_OWNED_FIELD_NAMES and admits it back only through dirtyEditPayload (edit only, objectui#12205); DrawerForm.tsx and writePayload.ts have the same shape. By code reading, every New form that shows owner_id (Account, Opportunity, Contact, Lead) drops the pick; only Lead was measured. · Seam: spec:Field.system owner_id (system: true, readonly: false, documented as reassignable) → renderer:plugin-form ModalForm create payload · Blocks nothing this card owns: the card's own defect is fixed, and this was the card's "manager can still pick another owner" check, which never worked on create. Seat decides whether the card waits (Blocked-by) or closes. · dedupe words: create form owner dropped; owner_id stripped on create; sanitizeFormData owner create; pick owner new record ignored; REASSIGNABLE_OWNER_FIELD create","carrier: none (承接者:无) · noted in PR #2099 Acceptance notes, not filed: lead.view.ts
formViews.advanced_conditional(a showcase form, not the create dialog) still marks owner_idrequired: truewhen status is contacted or qualified. Outside this card's file surface; I did not measure whether any console entry point opens that form."]
}
objectstack-fleet commentedon Oct 14, 2026
Review of PR #2099 (head
ca836d0d) against the os-dev report on this card (R85, seatrepo:hotcrm,session_01P8mZuzuyZAHSrgQKmv6ASv): REWORK, round 1 (sent to the same dev, same claim).main,Fixes #2097andClause-②: noat the top. The net diff is 3 files (+26 −4): the create form'sowner_identry, the pin intest/ownership-model.test.ts(whose reverse check went red as predicted and was restored by blob hash), and ahotcrm: patchchangeset.Fixesstands: the card's defect (a rep cannot save without picking themselves) is fixed here.owner_id, so the record is saved under the creator objectui#12359 (the create payload stripsowner_id; edit-only admission since objectui#12205). The comment is to say only what holds today.mainmoved toe094ec71(PR fix(views): zh-CN Leads and Opportunities column headers follow the field labels #2098, same file, columns block). The branch is to mergemainand re-read the ratchet, build, test and validate on the merged head, because the merge queue re-runs no build on the combined tree.6b5c8121added a 35,749-line dev-server temp file andb5237b67removed it (the net diff is clean; the squash carries only the net). Its model-namedCo-Authored-By:trailer is not rewritten (os-dev rule). Thegit fetch --unshallowin the shared hotcrm clone added history only.formViews.advanced_conditionalstill requiresowner_idfor contacted / qualified (PR Acceptance notes).Generated by Claude Code
objectstack-fleet commentedon Oct 14, 2026
os-dev-report
{
"issue": 2097,
"status": "done",
"round": "REWORK round 1",
"branch": "claude/issue-2097-lead-owner-not-required",
"pr": "#2099",
"head": "c1977c96",
"session": "session_01P8mZuzuyZAHSrgQKmv6ASv (this run is a subagent; the session id is the parent seat's)",
"premise_still_valid": true,
"summary": "Rework round 1, both items done. (1) 43a0b4c rewords the owner_id comment in the src/sales/views/lead.view.ts create form so that every claim in it is true today: a rep's new lead is theirs without a pick,
requiredblocked that save client-side, and a pick of another owner on create is currently dropped by the console (objectui#12359). The sentence "a manager can still choose another owner" is gone. Nothing else in the file changed (3 insertions, 2 deletions, all comment lines). I re-read the test comment in test/ownership-model.test.ts and the changeset: neither says a manager can pick another owner. The changeset only says "Lead Owner stays on the form", which is true. (2) c1977c9 merges origin/main e094ec7 (#2098 / #2094, the columns block of the same lead.view.ts) with a merge commit, no rebase and no force-push. The merge was clean, and the net diff against origin/main is still 3 files. All four requested gates are green on c1977c9, the combined tree. The PR body was not edited. ItsFixes #2097line and the round-0 notes stand as written.","tests": "All on HEAD c1977c9 (the merge of origin/main e094ec7). pnpm hygiene:tokens: "✓ source token ratchet clean", exit 0. pnpm validate: "✓ Validation passed (3064ms)", exit 0, lock "VERDICT command-exit 0". pnpm build: "✓ Build complete (2898ms)", exit 0, lock "VERDICT command-exit 0". pnpm test: "Test Files 179 passed (179)", "Tests 3660 passed | 1 skipped (3661)", lock "VERDICT command-exit 0", held the lock 8m34s (single command, no
;batch this time). Round 0 evidence (console measurements, the reverse check on the pin) still applies: this round changed only comment lines in lead.view.ts, plus the merged #2098 columns block. Not re-run this round, because they were not requested and the final diff did not touch their scope: typecheck, lint, lint:i18n-gate, hygiene (all green at ca836d0). NOT MEASURED: Playwright e2e (CI's).","mcp_calls": "0 — no MCP GitHub tool was called",
"api_writes": "1 — this os-dev-report comment via scripts/pm/post-stamped.mjs --comment=2097 (fleet relay). git push (1 this round, carrying 43a0b4c + c1977c9) is not REST. PR body not edited; zero label writes.",
"gates": {
"head": "c1977c96",
"pnpm hygiene:tokens": "exit 0 · ✓ source token ratchet clean",
"pnpm validate": "exit 0 · ✓ Validation passed",
"pnpm build": "exit 0 · ✓ Build complete",
"pnpm test": "exit 0 · Test Files 179 passed (179) · Tests 3660 passed | 1 skipped (3661)",
"pnpm typecheck": "not re-run this round (exit 0 at ca836d0)",
"pnpm lint": "not re-run this round (exit 0 at ca836d0)",
"pnpm lint:i18n-gate": "not re-run this round (exit 0 at ca836d0)",
"pnpm hygiene": "not re-run this round (exit 0 at ca836d0)",
"playwright e2e": "NOT MEASURED (CI)"
},
"line_budget": "net diff origin/main...HEAD (c1977c9): 3 files, +27 -4. src/sales/views/lead.view.ts +6 -4 (the form entry and its comment only). test/ownership-model.test.ts +16. .changeset/2097-lead-owner-not-required.md +5. No skills/** touched.",
"files_changed": [
"src/sales/views/lead.view.ts",
"test/ownership-model.test.ts",
".changeset/2097-lead-owner-not-required.md"
],
"deviations": [
"Carried from round 0, unchanged: commit 6b5c812 accidentally added the dev-server compile file
objectstack.config.bundled_4tw0mek20uf.mjs, and b5237b6 removes it, so the net diff does not include it. 6b5c812 also carries a harnessCo-Authored-Bytrailer that names the model. Neither was rewritten.","Round 1: none. 43a0b4c and the merge commit c1977c9 use only the model-free trailer pair. I used a merge commit, not a rebase, and did not force-push. The worktree was recreated from the existing branch at ca836d0 and removed after the gates passed."
],
"open_questions": [],
"out_of_scope_findings": [
"No new findings this round. The round-0 manager-pick finding has been filed upstream by the seat as objectstack-ai/objectui#12359, and the lead.view.ts comment now names it."
]
}
objectstack-fleet commentedon Oct 14, 2026
Review of PR #2099 round 1 (head
c1977c96) against the os-dev report6164367780(R85, seatrepo:hotcrm,session_01P8mZuzuyZAHSrgQKmv6ASv): ACCEPT.6164200283) are closed, as read on the PR:owner_idcomment in the create form now states only what holds: a rep's new lead is theirs without a pick;requiredblocked that save client-side; a pick of another owner on create is currently dropped by the console (plugin-form: a New form offers Owner, shows the picked user, and then drops the pick — the create payload stripsowner_id, so the record is saved under the creator objectui#12359).maine094ec71with a merge commit (no rebase, no force-push);mainis an ancestor of the head.main: 3 files, +27 −4: the form entry and its comment, the pin intest/ownership-model.test.ts, and thehotcrm: patchchangeset. Body stillFixes #2097/Clause-②: no, with no other closing keyword. Not a governed path.hygiene:tokens,validate,buildandtest(179 files, 3,660 passed) are green onc1977c96itself, the merge ofmain. typecheck / lint / lint:i18n-gate / hygiene were green atca836d0dand this round changed comment lines only. CI onc1977c96: every check green exceptBuild and Test (22.x), still running at this review; landing waits for it.Generated by Claude Code