refactor(signals): store node reads select by readerSeesCommitted (Rule 1, one implementation)

[ryansolid]

Stacked on #3515. Step 2 of DESIGN-CONSOLIDATION move 3b.

What moved

nodeValue — the store's untracked node-read selection — restated core read()'s committed-vs-staged arm by hand (inOwnerContext, a stale-of-foreign foreignHold twin, the HELD-truth mask). It now calls readerSeesCommitted (from #3515) with core's context selection (readerContext(): a root reads as its parent computed, context persists under untrack) behind the two store-only tunnels that stay in the store: latest() reaching this path for store keys (#3075) and authoritativeServe() (truth authors: the projection derive's draft, write-override).

Tracked store reads already went through readNode → core read() → readerSeesCommitted, so after this every store node read — tracked or untracked — selects by one Rule 1 implementation.

Not zero-change: the twin was missing a precondition

A stale reader (render effect) whose untracked read is of a store key held by a foreign action was served committed (correct, A15 / A26) but never recorded for replay at the commit (heldFromStale's _gatedSubs contract, which the signal path performs). The effect stayed on the pre-action value permanently after the action settled. Signal twin: [0, 0, 1]; store on next: [0, 0].

Pinned as S4 in posture-store-parity.test.ts (signal and store now both [0, 0, 1]; the store pin fails on next).

This is exactly the pattern the consolidation is for — one rule, N sites, one site missing a precondition — and it was invisible to the suite, the oracles and the 621-cell matrix (no reader kind there is "untracked inside a stale reader"). Found only by handing the store the shared predicate and asking what changed.

Remaining gaps, pinned / noted

• Backing twin (step 3): a key with no node goes through pendingBackingVisible / heldFromReader — same committed value, but there's no node to record the reader on, so no replay. Pinned it.fails beside S4. Rule 1's backing-level form is the next item.
• Lane arm: laneReadsCommitted (Existing and newly revealed latest readers display different values #3460) now also applies to untracked store node reads as it did to tracked ones. No matrix cell or test distinguishes; the matrix's foreignLane × untracked store cells are identical.

Gate

check	result
signals suite	2748 pass, 1 expected fail (the backing-twin pin)
visibility oracles + 621-cell posture matrix	byte-identical to next
fuzzer seed 3289 × 1000	994 / 0 / 6 — same as next (signals-only, so expected)
solid-js / web	623 pass / same 12 pre-existing server-diagnostics failures as origin/next
size	minified store −5 B, full bundle brotli −44 B vs #3515; no caps exceeded

Made with Cursor

[changeset-bot]

🦋 Changeset detected

Latest commit: cfd4dc0

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 11 packages

Name	Type
@solidjs/signals	Patch
test-integration	Patch
@solidjs/web	Patch
@solidjs/babel-plugin	Patch
@solidjs/compiler	Patch
@solidjs/diagnostics	Patch
@solidjs/element	Patch
@solidjs/h	Patch
@solidjs/html	Patch
solid-js	Patch
@solidjs/universal	Patch

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR

[ryansolid]

Superseded by #3523, which now targets next directly and carries this PR's commit (rebased) along with the rest of the stack. Closing so the move-3b work is one review rather than five.

— Claude via Cursor