fix(signals): optimistic settle verdicts — combined isPending on the child lane, onSettled after the revert re-derive (#3409, #3411)

[ryansolid]

Fixes #3409, fixes #3411 — the two optimistic-bookkeeping issues from gabbev's latest batch. Both are small, and both were reproduced core-level (manual clock) before fixing; the pins are in optimistic-settle-verdicts.test.ts and fail on next.

#3409 — combined isPending lags the individual probes

isPending(() => [a(), b()]) over two async siblings of an optimistic value reported false while isPending(a) and isPending(b) both reported true, and flipped to true only once a landed.

Traced: the combined effect computes true at the write, but its run is withheld — it ends up on the held value lane. Sequence in assignOrMergeLane: ps(a)'s child lane is assigned → a's parent-lane notification keeps the child (the parent/child carve-out) → ps(b)'s child lane is a sibling, so the two child lanes merge → b's parent-lane notification then hits the "subscriber's lane was merged → replace with the source lane" shortcut, which skips the parent/child check and puts the effect on the held parent. fxA/fxB each have a single child lane and never reach that path.

Fix: delete the shortcut. A merged lane is followed to its root (findLane) like any other and goes through the same parent/child logic — the root is a child of the source root, so the effect keeps it. −33 B on the core floor.

The pin runs through the revert too: there the three indicators clear together when the held reload commits, which matches the plain-signal baseline.

#3411 — onSettled reads Value: 0, Copy: 1 during the revert

The settle drops the override in the commit pass but only enqueues the reverted subscribers (_resolveOptimistic → insertSubs, likewise the contested/gated replays and store clears); the pass after re-derives them, and reads do not pull (prepareComputed(el, false)). An unowned onSettled is a one-shot in the commit pass's user phase, so it read value already reverted next to copy still holding 1.

Fix: the unowned fire waits for the heap to drain — while dirtyQueue has work it re-enqueues itself (run swaps the queue array, so the re-enqueue lands in the next pass; enqueue keeps the drain alive). That is what settled means: no derivation outstanding. Also covers gabbev's no-action variant (ambient optimistic write reverting at the end of its own flush), pinned separately.

I first tried the other lever — re-running the heap inside finalizePureQueue after _resolveOptimistic so the commit pass itself is consistent. It breaks three things that rely on the next-pass re-derive: the lane frame of an optimistic write completing in its own flush (applied by cleanupCompletedLanes; [0, 2, 0] became [0, 0]), the store optimistic-layer clears (the #3164 test produced a torn RUNTIME ERROR frame), and a gated replay. The internals doc records why the observer-side fix is the right one.

Verification

• Full monorepo turbo run test serially, rebased on current next (includes fix(signals): defer a transaction-owned node's committed children to its commit (#3404) #3405): 23/23 tasks green.
• Core floor 23,091 → 23,058 (−33 B); all ten size caps pass, no bumps.
• src/ typechecks clean (the 26 test-file type errors are pre-existing on next).
• Changeset + internals doc bullets (§2 lanes, §3 transitions).

Co-authored with Claude via Cursor.

[changeset-bot]

🦋 Changeset detected

Latest commit: 9fa294f

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

[coveralls]

Coverage Report for CI Build 34830828074

Coverage remained the same at 71.842%

Details

• Coverage remained the same as the base build.
• Patch coverage: No coverable lines changed in this PR.
• No coverage regressions found.

Uncovered Changes

No uncovered changes found.

Coverage Regressions

No coverage regressions found.

---

Coverage Stats

Relevant Lines:	1007
Covered Lines:	772
Line Coverage:	76.66%
Relevant Branches:	790
Covered Branches:	519
Branch Coverage:	65.7%
Branches in Coverage %:	Yes
Coverage Strength:	15.03 hits per line

---

💛 - Coveralls

[codspeed]

Merging this PR will not alter performance

✅ 160 untouched benchmarks

---

_{Comparing fix/optimistic-settle-verdicts (9fa294f) with next (1643d2a)}