hydration: readShallow via ownKeys, store adapters ride the trace chunk, SC-only installs moved off hydrate()

.changeset/server-component-hydration-installs-from-frames.md

@@ -0,0 +1,8 @@
+---
+"solid-js": patch
+"@solidjs/web": patch
+---
+
+The server-component half of hydration installs from `@solidjs/web/frames`' `installServerComponents()`, not from `hydrate()`: `sharedConfig.holdBoundary` (the client hold on adopted markup counted as a pending boundary), `sharedConfig.hydrateWindow` (the claim window an adopted occurrence re-enters hydration through), fragment ownership by rendering (the `_$HY.fa` term of the reveal policy) and the ledger's published answer `_$HY.fr` on the solid side; the declared claim roots (`sharedConfig.claimRoots`) and the frame-region exclusion of the root sweep on the DOM runtime's side. A hydrating page without server components carries none of it — −686 B minified / ≈ −205…−243 B brotli on the hydrating apps (this returns the bytes the frames A0 correctness pass added to every hydrating page); a server-component page pays the two installers' glue (≈ +220 B minified).
+
+Public surface (`@internal`): `enableServerComponentHydration()` on `solid-js` (runtime) and `solid-js/internal` (typed), a no-op on the server entry; `installServerComponentHydration()` on `@solidjs/web`, which installs both halves — the frames client calls it where it installs its reveal hook.

.changeset/store-hydration-adapters-ride-the-trace-chunk.md

@@ -0,0 +1,7 @@
+---
+"solid-js": patch
+---
+
+The store hydration adapters (`createStore(fn)` / `createProjection` / `createOptimisticStore(fn)` under hydration: snapshot adoption, the parked patch backlog, the hybrid handoff) move to their own source module, and `solid-js/internal/container-trace` — the frames traces tier's materializer, a lazy chunk — bundles its own copy of it instead of reading the eager slot `enableHydration()` fills through `withStoreHydration`. Reading the slot from the lazy chunk pinned every adapter into the entry chunk of a server-component page that never creates a client store (an app bundler keeps a module eager whenever an eager module imports it); the copy is inert by construction — the adapters declare no module-level state, and every piece of shared state is read back from the one `solid-js` instance. −2,550 B minified / ≈ −575 B brotli on the compiled server-component page; 0 on pages whose root pass creates a derived store (they keep the eager install); the lazy `trace.js` chunk grows ≈ 0.5 KB brotli.
+
+Public surface (`@internal`, stripped from the declarations): `withStoreHydration` is removed from the `solid-js` client and server entries (its only consumer was the trace entry); the client entry gains the runtime exports the adapter copy reads — `readSerializedOrCompute`, `subFetch`, `readHydratedValue`, `wrapFirstYield`, `adoptedAnswerStream`, `withHydrationGate`, `onHydrationEnd`, `noHydrationId`, `markTopLevelSnapshotScope`, `hasLoadingWindow`, `isAsyncIterable`, `syncThenable`, `UNASKED` — mirrored on the server entry as inert stubs for export parity.

.changeset/web-read-shallow-own-keys.md

@@ -0,0 +1,5 @@
+---
+"@solidjs/web": patch
+---
+
+`readShallow` (the compiler's one-layer tracked read of an object-valued `class`/`style`) reads a proxy's keys with `Reflect.ownKeys` directly — what `sourceKeys(value, SOURCE_PROXY)` resolves to by definition — instead of through `sourceKeys`, whose runtime kind argument kept the merge/omit view walkers of `@solidjs/signals` in every compiled page with a dynamic `class` or `style`, reached by nothing. Identical behaviour; −1,165 B minified / ≈ −350 B brotli on the compiled server-component page (a page with an element spread keeps the walkers through `spread()`).

documentation/plans/frames-savings-pass.md

@@ -491,6 +491,16 @@ Pages (br; the whole page, lazy chunks not counted):
 | compiled hydrating        | 30,943 (at its 30.93 cap) | **≈ 30,990** (the same +≈ 45)                                                                                                                                    | 31,075 (+52 — Phase A's)                                                                                    | unchanged                            | +45                  | **cap raise needed — the maintainer's**                            |
 | frames eager              | 13,770                    | ≈ 7,600 (7.8 reading; the C6 gate is 8.5 before the residual R / D cuts land) / ≈ 8,150 (8.0)                                                                    | **13,083** (40,000 min; **−704 br**; chunks `trace.js` 8,176 / `regions.js` 805 / `assets.js` 783 br, lazy) | **≈ 7,250 / ≈ 7,800**                | **−6.5 KB (−47 %)**  | ≈ 10,500                                                           |
 
+The hydration pass (2026-10-07, `size/hydration-pass`;
+`hydration-split-measured.md`) returned the Phase A bytes to the plain
+hydrating pages and more — `app: hydrating (no stores)` 17,894 → 17,689 br
+and `compiled hydrating` 31,155 → 30,991 against `next` @ `d231b9911` — by
+moving the server-component half of hydration (the hold, the claim window,
+fragment ownership, `_$HY.fr`, the claim-roots walk and the frame exclusion
+of the root sweep) behind `installServerComponents()`, and took `page:
+compiled base SC` 35,146 → 34,321 br with the store hydration adapters as
+the trace chunk's own copy and `readShallow` off the view walkers.
+
 ### 4.1 Measured (2026-10-06)
 
 The Phase D estimates above, turned into numbers by a measurement pass on