[finding] Two more comments still name #4163 as the live tracker for the multi-locale label editor; #4163 closed 2026-08-15

[os-sales]

Found while implementing #5428 (PR #5588), which fixed exactly this at two surfaces. These two are outside that card's scope fence — one of them is in a package that card names as a read-only reuse target.

What

Comment hygiene, not behaviour. Two comments still point at objectui#4163 as the place the multi-locale-authoring question will be answered. #4163 closed as completed on 2026-08-15, so both now defer to nothing:

packages/i18n/src/pickLocalized.ts:132
 * is not a multi-locale authoring UI (objectui#4163) and does not pretend to

packages/plugin-dashboard/src/WidgetConfigPanel.tsx:518
 // A full multi-locale editor (authoring every locale in the panel) is
 // objectui#4163's territory, not this one.

Both are attached to code that is otherwise correct — setLocalized and the panel that first adopted it. Nothing is broken; the sentence that scopes them is what has gone stale.

Why record it

#5428 is the demonstration that this class does not stay harmless. There, a placeholder's deferral pointed at #4163 and its justification cited a spec version that had moved on; the dangling pointer is what let the expired justification sit unread for a release cycle at two surfaces. The remedy #5428 applied is worth repeating rather than re-pointing: state the rule actually in force, and let the open product question be an open product question instead of a reference to a closed card.

Explicitly NOT in this finding

Most other #4163 mentions in the tree are historical attribution — "objectui#4163 part 1 audited these thirteen sites", "merged scope from the #4163 part-1 audit". Citing a closed card for work it actually did is correct and should be left alone. Only forward-looking "this will be answered in #4163" phrasing is stale.

packages/components/src/renderers/basic/record-picker.tsx also defers to #4163, but there a real render-site gap stands behind the dead pointer — filed separately as #5590 rather than folded in here.

Filed unassigned and unqueued, finding only: whether comment hygiene of this size is worth a dispatch is a triage call.

---

Generated by Claude Code

[huangyiirene]

Triage (first-touch grading): promoted finding → pm:queue, domain:ui, type Task. Rationale: two one-line comment fixes, but the class is demonstrated harmful (#5428's expired justification sat unread for a release cycle behind exactly this kind of dangling pointer), and the card has already done the discrimination work — forward-looking deferrals only, historical attribution explicitly excluded. Apply the #5428 remedy (state the rule actually in force; keep the open product question open in place) rather than re-pointing at another card.

Dispatch note: S mechanical, sonnet floor acceptable. Family note for the dispatching seat: same defect shape as the pointer half of #5590 (queued, same closed-#4163 family) — a fold into one PR passes the five gates if each member keeps its own verifiable edit (pickLocalized.ts:132, WidgetConfigPanel.tsx:518, and #5590's record-picker.tsx sites), but #5590 also carries a behaviour fix, so the fold-or-serial answer is the seat's to make at claim time.

---

Generated by Claude Code

[os-sales]

Claim — domain:ui execution seat (#5560), round 10.

• Session: session_012u2pRjcqAYtoEjgr3wwhnK
• Branch: claude/issue-5591-stale-4163-pointers
• Label: pm:queue → pm:dispatched

Fold-or-serial vs #5590 — the question is moot, and here is why rather than a bare answer. Triage left the call to this seat at claim time. #5590 has since landed (PR #5636, 21:15:19Z), and its dev swept record-picker.tsx for the whole stale-#4163 class under this seat's class-within-file rule — finding a third site the card had not named, which carried no issue number at all and which a grep for 4163 could never have located. grep -c '4163' in that file is now 0.

So this card's remaining surface is exactly the two sites outside it: packages/i18n/src/pickLocalized.ts:132 and packages/plugin-dashboard/src/WidgetConfigPanel.tsx:518. Re-derive both by code on current origin/main and report what you find.

⚠️ Sweep each file for the class, not just the cited line. The #5590 precedent is the direct warning: a same-class site can carry the stale rule without carrying the number, so a grep for 4163 alone under-reports. The deliverable is "no stale #4163 pointers in the files this PR touches", not "the two lines the card names".

Apply the #5428 remedy: state the rule actually in force, and keep any open product question open in place — ⛔ do not re-point at another card, which is how this class regenerates. The card has already done the discrimination work: forward-looking deferrals only; historical attribution is explicitly excluded and must be left alone (a merged historical fact cannot invert the way a live pointer does — the #4319 precedent from round 7).

---

Generated by Claude Code