From 058015098061944b3c653a2257da9162b1606226 Mon Sep 17 00:00:00 2001 From: Shevchik Igor Date: Sat, 29 Aug 2026 05:09:22 +0000 Subject: [PATCH] chore(sync): reconcile the #509 ledger entries with their merged PR MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit `a494a97d`, `3d2de0ce` and `ae243116` were written while their port was still open, so all three carried `pending-merge` in both `pr` and `b24ui_sha`. #509 squash-merged as `7203420b`; this points them at it. Bookkeeping only — no ledger entry is added, the cursor does not move, and no decision changes. --- .sync/nuxt-ui.json | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) diff --git a/.sync/nuxt-ui.json b/.sync/nuxt-ui.json index 45d49f6ea..6b1e4aa77 100644 --- a/.sync/nuxt-ui.json +++ b/.sync/nuxt-ui.json @@ -1669,20 +1669,20 @@ "summary": "fix(Checkbox/RadioGroup): use `border-default` on `card` and `table` variants (nuxt/ui #6884) — NO-OP: the line it edits is written in different tokens here. Upstream swaps one token across five slots, border-muted -> border-default on the card and table variants of Checkbox, CheckboxGroup and RadioGroup, with 168 snapshot lines following. Neither token is on our side of that line: all five fork equivalents read `border border-(--ui-color-design-outline-na-stroke)`, the Air design system's outline stroke, annotated @memo style-outline-no-accent in checkbox.ts — checked individually at checkbox.ts:63, checkbox-group.ts:48,56 and radio-group.ts:69,77. There is no border-muted to replace and no border-default to replace it with. The ${hover}bg-elevated/50 fragment in the same string is not ours either: it arrived upstream in 07f3fe8d as part of the hover/checked/focus palette refresh, which this fork DECLINED by maintainer decision when that commit was ported (#464), our card and table carrying their own checked and hover treatment in air tokens. So this commit adjusts a line that on our side was already deliberately different. Stated precisely rather than as 'we don't use those tokens': we do use border-default elsewhere (dropdown-menu.ts and sidebar.ts x4) and border-muted in prose/prompt.ts — the tokens exist here, they are simply not what these five slots are written in. Nothing to port and nothing deferred: taking it would mean first adopting the palette refresh this fork chose not to." }, "a494a97daf9d9dffe9dc38c4755c8964e666ed46": { - "pr": "pending-merge", - "b24ui_sha": "pending-merge", + "pr": "509", + "b24ui_sha": "7203420bd074d1081d4d540133e3e98bdc96da86", "decision": "port", "summary": "fix(Form): omit method attribute on nested forms (nuxt/ui #6888) — PORT, verbatim, one line. A nested Form renders as a div rather than a form (:is=\"parentBus ? 'div' : 'form'\") because a form inside a form is invalid HTML, but method was still set unconditionally, so the nested case emitted
— a form attribute on an element that is not a form. Applies verbatim: our Form.vue has the same pre-image, with the root already conditional and :name already parentBus ? undefined : props.name, so method is simply the one attribute missed when the nested case was introduced; nothing about the line involves ui/b24ui. No snapshot moved — Form.spec.ts does not render a nested form, so the attribute change is not witnessed by the suite. Stated rather than glossed over: the fix is correct by inspection of the template and upstream's reasoning, not by a test that would have caught it." }, "3d2de0ce790f3fd90a95447b7145bf25c9ee4b3a": { - "pr": "pending-merge", - "b24ui_sha": "pending-merge", + "pr": "509", + "b24ui_sha": "7203420bd074d1081d4d540133e3e98bdc96da86", "decision": "port", "summary": "fix(Slider): emit a number for a single thumb (nuxt/ui #6890) — PORT, verbatim, against Range. Upstream's Slider is this fork's Range (§1); third commit in that family after f3c2ac21 (#431) and d6c3802a (#466). Passing emits to useForwardProps forwarded SliderRoot's own update:modelValue straight through, IN ADDITION to the one the component derives from its defineModel: the event fired twice, and the forwarded copy carried Reka's raw payload — an array — so a single-thumb slider emitted [71] where the caller passed and expects 71. Dropping the argument leaves the component's own handling as the sole emitter. Verbatim here: our Range.vue:74 carried the identical emits argument, and Range.spec.ts carried the identical pre-image expectation { 'update:modelValue': [[1], [1]] } — the doubled emit written down as though it were correct. Ported upstream's corrected expectation and both new cases. Mutation-checked rather than assumed: restoring the emits argument with the tests kept turns all six red — the corrected case plus the two new ones, across both the nuxt and vue projects." }, "ae24311663cc8447c6e44b6045087c94d3ae4c71": { - "pr": "pending-merge", - "b24ui_sha": "pending-merge", + "pr": "509", + "b24ui_sha": "7203420bd074d1081d4d540133e3e98bdc96da86", "decision": "skip", "summary": "docs: refactor with `nuxt-agent-discovery` module (nuxt/ui #6883) — SKIP: it refactors infrastructure this fork deliberately does not have. 33 files, +474/-1914, replacing the hand-rolled agent-discovery surface upstream built in 9ef3ee39 with a nuxt-agent-discovery module and deleting its own implementation: .well-known/api-catalog.get.ts (-63), .well-known/mcp/server-card.json.get.ts (-44), raw/[...slug].md.get.ts (-64), raw/index.md.get.ts (-79), sitemap.md.get.ts (-47), sitemap.xml.get.ts (-29), markdownNegotiation.ts (-318), middleware/markdown.ts (-66), error.ts (-58), extractSections.ts (-61), useCanonical.ts (-24) and robots.txt (-40), replaced by server/plugins/agent-discovery.ts (+58) plus a dependency. 9ef3ee39 — the commit that built this surface — is recorded one entry earlier as a skip on maintainer decision, because our agent surface is our own and larger (docs/server/mcp/ carries 27 files against upstream's smaller set), so a refactor of code we never took has nothing to apply to. And taking it would delete working code written on purpose: the two .well-known routes upstream removes here are the ones this fork ADDED in #492, precisely because nuxt.config.ts advertised them and nothing served them; our raw/* and sitemap.* routes are likewise our own and in use, with raw/**.md being what skills/b24-ui-nuxt/references/components.md links to and both sitemaps prerendered into the static build. Adopting nuxt-agent-discovery remains a live option, but it is a maintainer decision about our own architecture rather than a port — it would mean deleting six working routes and four utilities for a third-party module whose output would then need re-verifying against GitHub Pages, where server routes do not run at all." }