docs(sync): record that upstream's Slider is this fork's Range

[IgorShevchik]

Added after I answered a sync check wrongly.

What happened

The queue holds nuxt/ui@f3c2ac21, fix(Slider): bind form aria attributes on thumbs instead of root. I reported it as a no-op, on these grounds:

there is no Slider component anywhere in src/, and 249 ledger entries never mention one

Both statements are true. The conclusion was wrong: the search was by name, and the name is the one thing that changed.

src/runtime/components/Range.vue is upstream's Slider. It wraps the same SliderRoot / SliderTrack / SliderRange / SliderThumb, has the same root/track/range/thumb slots and the same color/size/orientation/disabled variants, its theme is src/theme/range.ts, its page is range.md — and its props line is upstream's verbatim:

export interface RangeProps extends Pick<SliderRootProps, 'name' | 'disabled' | 'inverted' | 'min' | 'max' | 'step' | 'minStepsBetweenThumbs'> {

The fix applies here in full. It is ported separately, in queue order.

Why there was nothing to catch it

.sync/ had no component-name map at all. icon-map.json and color-map.json cover tokens; nothing covered wrappers. So a name-based search returning nothing had no counterweight, and "the fork does not have this component" reads like a recorded decision rather than an artefact of the query.

§1 now carries:

• the rename, as a row in the mechanical-rewrites table;
• why grepping an upstream component name is the wrong move — only the b24ui wrapper is renamed. Every reka-ui export keeps its own name inside the file, so a diff touching SliderThumb needs no rewriting at all. Search the wrapped primitive or the props, not the wrapper;
• the full comparison, so the next reader does not have to trust a single example.

The comparison

Both trees hold 180 components. 19 exist only upstream, 19 only here, and none of those 38 is a second rename — checked pairwise rather than assumed.

• upstream only — AuthForm, BlogPost, BlogPosts, Carousel, ChangelogVersion, ChangelogVersions, Icon, Marquee, PageAnchors, PageCTA, PageHero, PageLogos, PricingPlan, PricingPlans, PricingTable, Tree, content/ContentNavigation, prose/CodeTree, prose/Icon
• b24ui only — Advice, Countdown, DescriptionList, ModalDialogClose, Navbar, NavbarDivider, NavbarSection, NavbarSpacer, PageCardGroup, SidebarBody, SidebarFooter, SidebarHeader, SidebarHeading, SidebarLayout, SidebarSection, SidebarSpacer, TableWrapper, prose/H5, prose/H6

The note ships the command to re-derive both lists, since a list in prose is a snapshot. I ran it against the note's own figures before committing: 180 vs 180, 40 differing names.

Deliberately untested

Worth being explicit, since this repo's habit is to guard every new rule. A test here could only assert that our side of a map exists — Range.vue does exist, so such a guard would have passed while I was drawing the wrong conclusion. It would add the appearance of coverage without the substance. The rule is documentation, and says so.

Verify

lint and test/utils (415 tests, 22 files). Markdown only; nothing executable changes.

---

Generated by Claude Code