fix(signals): a render effect that stops reading a pending memo releases the memo's source (fuzzer #3446 P1)

[ryansolid]

Stacked on #3485 (the posture matrix cell this fixes); merges cleanly after it.

Bug

A render effect that directly observed a pending async memo, then re-ran without reading it (a show() gate closed), kept the memo's source held — forever when the flight never lands:

const m = createMemo(() => { s(); return new Promise(() => {}); });
createRenderEffect(() => (show() ? m() : "hidden"), …);
flush(); setS(1); flush();     // s held by m's flight — correct
setShow(false); flush();       // effect re-runs "hidden", no longer reads m
// before: s() === 0 and isPending(s) === true, forever

Found by the semantic fuzzer (#3446, law P1: "ordinary writes publish after a drain when no visible reader still needs an unresolved answer") — 21 of 1,000 cases in one symptom group, case 854. Reproduced as the posture matrix's effect × gatedAway cell; the memo-between variant (memo × gatedAway) already released, which narrowed it to the direct-effect reporter.

Cause

Not the liveness predicate: reporterBlocksSource already judged the re-run effect dead (no pending status, no pending sources, deps without the memo). Nobody re-asked: the flush that re-ran the effect had no active transaction, and transitionComplete only judges the active one. wokenTransitions' own doc comment names the class — "a reporter that stops counting for another reason … no check comes" — with two sites: disposal (#3372) and boundary reset (#3375). This is the third.

Fix

recompute's tail: a pending reporter (render effect) that recovers without its flight landing — wasPending && !(STATUS_PENDING) after the pass — wakes the transaction it reported to (wokenTransitions.push(t); schedule()), skipped under that transaction's own flush (the landing that recovers a reader there is judged by that flush). Reporters are stamped, so no unstamped fallback. +100 B minified (24,478 → 24,578); +createStore brotli cap +1 B.

Fuzzer, same campaign against the fix: 21 → 4 failures, policy unchanged (12; those translate to plain Solid that publishes synchronously — the fuzzer's wait-budget model).

Remaining, pinned it.fails (spec O3, remaining form)

The 4 are one shape: gate and write in the same flush (case 21; case 79 multi-step). The effect is notified pending by the write (registering as reporter) and dirtied by the gate in the same flush; the verdict runs after the pure phase, before effects, so the effect still looks live, the transaction parks, and the effect's run — the one that would prove it dead — is stashed with the transaction. The hold keeps the run that would release it. That's the effect-phase parking question (#3407), not a fifth predicate arm; left for the consolidation.

Tests

posture-born-held-and-observation.test.ts: P1 pin flipped from it.fails to it; same-flush form added as it.fails. signals 2,495 + 1 expected fail; solid 620; web 823.

[changeset-bot]

🦋 Changeset detected

Latest commit: 62b0a22

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 35067907246

Coverage remained the same at 71.46%

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:	1018
Covered Lines:	772
Line Coverage:	75.83%
Relevant Branches:	790
Covered Branches:	520
Branch Coverage:	65.82%
Branches in Coverage %:	Yes
Coverage Strength:	14.94 hits per line

---

💛 - Coveralls

[codspeed]

Merging this PR will not alter performance

✅ 175 untouched benchmarks

---

_{Comparing fix/gated-reporter-hold (62b0a22) with next (48463f5)¹}

Footnotes

1. No successful run was found on next (3a09d3f) during the generation of this report, so 48463f5 was used instead as the comparison base. There might be some changes unrelated to this pull request in this report. ↩