#1934 35682a73136e7483bfd22e14659e9a307155f296 Thanks @ashunar0! - feat(inertia): add infinite scroll with scroll()

Adds a scroll() helper that wraps a paginated page payload with the metadata
the Inertia client's <InfiniteScroll> adapter needs to keep loading more items
as the user scrolls:

users: scroll({
  data: await db.users.page(currentPage),
  currentPage,
  lastPage: 10,
  pageName: 'users_page',
  matchOn: 'id',
})

The value travels as-is; the renderer emits page.scrollProps[key] = { previousPage, nextPage, currentPage, pageName } on every response and opts the
prop into the merge protocol — defaulting to append, switched to prepend
when the client sends X-Inertia-Infinite-Scroll-Merge-Intent: prepend. The
optional matchOn is forwarded as page.matchPropsOn for client-side dedupe.
Mirrors Inertia::scroll(...) in inertia-laravel 3.x.

#1932 5318463bd4be3b6c7abb6f5c46a61bd1d5b36ca6 Thanks @ashunar0! - feat(inertia): add merge props with merge(), prepend(), and deepMerge()

Adds three helpers that mark a prop for client-side combination on the
next partial reload, instead of the default replace behavior:

• merge(data, { matchOn? }) — appends array items / shallow-spreads object keys.
• prepend(data, { matchOn? }) — same as merge() but inserts at the start.
• deepMerge(data, { matchOn? }) — recurses into nested arrays/objects (use for
  wrapper-shaped paginated props like { data: [...], meta: {...} }).

The value travels as-is; the renderer advertises which keys to combine via the
new page.mergeProps / page.prependProps / page.deepMergeProps arrays and
emits dot-paths on page.matchPropsOn for the client's dedupe logic.

import { deepMerge, inertia, merge, prepend } from '@hono/inertia'

app.use(inertia())

app.get('/feed', (c) =>
  c.render('Feed', {
    posts: merge(await db.posts.page(n), { matchOn: 'id' }),
    notifications: prepend(await fetchNotifications(), { matchOn: 'id' }),
    conversations: deepMerge(
      { data: await db.messages.page(n), meta: { nextCursor } },
      { matchOn: 'data.id' }
    ),
  })
)

The merge metadata is emitted on every response (initial + partial) so the
client knows which keys to combine on the next partial reload. Full page visits
always replace props entirely.

#1911 c5758a6e4b3ca44c8c03a06effc2e6c7134e4d25 Thanks @{! - feat(inertia): add deferred props with defer()

Adds a defer(resolver, group?) helper that defers a prop until after the
initial render. On the initial response the resolver is skipped and the
prop key is advertised via page.deferredProps[group]; the Inertia client
then issues one partial reload per group, at which point the resolver runs
and the value is sent down.

import { defer, inertia } from '@hono/inertia'

app.use(inertia())

app.get('/', (c) =>
  c.render('Dashboard', {
 id: 1 },                       // sent on initial response
    posts: defer(() => fetchPosts()),      // fetched after mount
    stats: defer(() => fetchStats(), 'secondary'),
  }),
)

Multiple deferred props that share a group are fetched together. The
default group is "default".