fix: router performance

[lazarv]

PR #372 ("enhanced router") introduced a regression where every <Route> in a layout caused the SSR worker to import the target page module and the browser to preload its chunk, even for non-matching siblings. On large layouts this dominated
request latency and forced every client chunk into the initial download. This PR removes that cost without giving up the correctness or DX of the new router.

The core idea is to keep live client references out of the Flight payload for routes that don't match. The file router now emits <Route componentId="…" componentLoader={() => Page} /> instead of <Route element={<Page/>} />. The live
reference only exists inside the closure body, so React's RSC encoder walks past it for non-matching siblings and never registers the chunk. Only the matching route calls componentLoader() and instantiates the page, which produces exactly
one client-reference registration per request. For non-matching routes we resolve the source-relative $id to the built chunk URL via clientReferenceMap (with a try/catch fallback to the source module so dev still works without the
.react-server/ build output) and pass the chunk id to the client.

On the client, ClientRouteRegistration builds a small LazyChunkComponent wrapper around the deferred chunk. It deliberately does not use React.lazy, because lazy always schedules a microtask before re-rendering and causes a one-frame
fallback flash even when the module is already in the __webpack_require__ cache. Instead the wrapper reads p.value synchronously on cache hits and falls through to React 19's use(p) hook only when the import is genuinely in flight. It
also patches .value/.status onto the import promise itself, since the prod polyfill in render-rsc.jsx does this on the server but Vite's dev __webpack_require__ does not. There is a load-bearing invariant: in lazy mode the wrapper is
only instantiated when the route is active, because Activity hidden subtrees still render and would otherwise eagerly fire the dynamic import for every sibling. The lazy render path also intentionally has no local Suspense boundary, so an
active route's suspension propagates to the navigation transition and React keeps the previous page visible until the new chunk resolves, instead of flashing a blank fallback.

The second half of this PR is an unrelated scroll-restoration bug surfaced by the new lazy navigation timing. ScrollRestoration initialised its lastY snapshot from window.scrollY at effect setup time, but on popstate the browser
carries the previous page's scroll position over because we set history.scrollRestoration = "manual". If a fast follow-up navigation ran the cleanup before any real scroll event refreshed the snapshot, the cleanup would write the previous
route's scroll value under the current route's key, silently corrupting saved positions. The fix introduces a module-level scrollObserved flag that the window scroll listener and every container scroll listener flip on first event, and the
cleanup only persists if a real scroll has been observed for that route. If nothing scrolled, the existing storage entry is correct and we leave it alone. This affects real users on any quick back/forward navigation, not just tests.

The scroll restoration test file was also rewritten to boot the dev server once via beforeAll (the previous per-test await server(...) was rebuilding the dev server before every test and dominating the suite duration — 63s down to 17s),
with a beforeEach that navigates to the fixture origin first and then clears sessionStorage, since storage is per-origin and clearing on about:blank is a no-op for the test origin.

Finally, the benchmark example was restructured into (rsc), (ssr), and (hybrid) route groups, and bench.mjs now exposes a --filter flag for local iteration plus a new set of hybrid benchmarks that exercise the
layout-with-many-client-siblings shape this PR is designed to make fast. Original benchmark URLs are unchanged (route groups are transparent), so the CI baseline comparison still works without modification.

[github-actions]

⚡ Benchmark Results

	PR d78da34	main 249497a
Config	50 connections, 10s/test	50 connections, 10s/test

Benchmark	Req/s	vs main	Avg Latency	vs main	P99 Latency	Throughput
minimal	1415	🟢 +197.4%	34.77 ms	🟢 -66.6%	70 ms	0.8 MB/s
small	1474	🟢 +194.4%	33.39 ms	🟢 -66.1%	59 ms	1.3 MB/s
medium	336	🟢 +58.6%	147.88 ms	🟢 -36.2%	215 ms	4.9 MB/s
large	31	🟢 +24.7%	1438.3 ms	🟢 -21.2%	2888 ms	3.1 MB/s
deep	933	🟢 +150.3%	52.96 ms	🟢 -60.2%	92 ms	3.1 MB/s
wide	41	🟢 +32.8%	1178.12 ms	🟢 -16.6%	2202 ms	2.2 MB/s
cached	3586	🟢 +13.6%	13.45 ms	🟢 -12.3%	30 ms	52.3 MB/s
client-min	516	🔴 -10.3%	95.77 ms	🔴 +11.0%	149 ms	2.6 MB/s
client-small	547	🔴 -3.7%	90.3 ms	🔴 +3.3%	137 ms	3.0 MB/s
client-med	407	🔴 -2.5%	121.97 ms	🔴 +2.5%	179 ms	7.8 MB/s
client-large	93	🟢 +7.2%	531.95 ms	🟢 -4.3%	995 ms	9.8 MB/s
client-deep	506	🔴 -3.0%	98.04 ms	🔴 +2.9%	141 ms	3.9 MB/s
client-wide	154	🟢 +11.1%	317.89 ms	🟢 -9.3%	538 ms	9.1 MB/s
static-json	9688	🔴 -11.3%	4.68 ms	🔴 +19.7%	14 ms	4.0 MB/s
static-js	9290	🔴 -15.5%	4.88 ms	🔴 +25.1%	15 ms	11.1 MB/s
404-miss	5553	🔴 -8.4%	8.45 ms	🔴 +10.7%	19 ms	0.7 MB/s
hybrid-min	505		98.12 ms		158 ms	2.8 MB/s
hybrid-small	497		99.97 ms		160 ms	3.3 MB/s
hybrid-medium	214		231.82 ms		359 ms	9.6 MB/s
hybrid-large	26		1701.65 ms		3338 ms	8.9 MB/s
hybrid-deep	380		130.41 ms		197 ms	5.6 MB/s
hybrid-wide	32		1371.84 ms		2166 ms	7.0 MB/s
hybrid-cached	2904		16.75 ms		33 ms	130.1 MB/s
hybrid-client-min	546		90.73 ms		133 ms	2.8 MB/s
hybrid-client-small	537		92.26 ms		136 ms	3.0 MB/s
hybrid-client-medium	397		124.26 ms		183 ms	7.6 MB/s
hybrid-client-large	91		538.59 ms		969 ms	9.6 MB/s
hybrid-client-deep	497		99.55 ms		151 ms	3.9 MB/s
hybrid-client-wide	147		334.62 ms		534 ms	8.7 MB/s

Legend

🟢 > 1% improvement | 🔴 > 1% regression | ⚪ within noise margin

Benchmarks run on GitHub Actions runners (shared infrastructure) — expect ~5% variance between runs. Consistent directional changes across multiple routes are more meaningful than any single number.

[codecov-commenter]

❌ 1 Tests Failed:

Tests completed	Failed	Passed	Skipped
849	1	848	19

View the top 2 failed test(s) by shortest run time

__test__/scroll-restoration.spec.mjs > scroll restoration: query-param-only change preserves scroll

Stack Traces | 2.51s run time

AssertionError: expected 0 to be greater than 500
 ❯ __test__/scroll-restoration.spec.mjs:174:19

__test__/scroll-restoration.spec.mjs > scroll restoration: back navigation restores scroll position

Stack Traces | 24.6s run time

AssertionError: expected 0 to be greater than 700
 ❯ __test__/scroll-restoration.spec.mjs:126:21

To view more test analytics, go to the Test Analytics Dashboard
_{📋 Got 3 mins? Take this short survey to help us improve Test Analytics.}