[GabbeV]

Summary

After #4498, an observed createMemo(() => 1 / value()) over createSignal(-0, { equals: Object.is }) stays at -Infinity after setValue(0), even though the signal commits +0. With an async action holding the write, the same equality shortcut also reports the write as not pending and loses the reader's replay at reveal. Both signed-zero directions and equals: false are affected.

Require Object.is as well as the existing strict comparison when identifying netted staging, in the fast read, strict slow read and verdict-routing paths. The test-only reader-seat table uses the same criterion. Opposite signed zeros remain distinct proposals; an accepted NaN rewrite retains its previous behavior because the strict comparison remains. User comparators, netting priority and allocation/state layout are unchanged.

Adds 22 regression/control cases: ordinary derivations, held plain reads versus latest, tracked isPending, both write locations around an action, strict reads, default equality and an accepted NaN rewrite, plus a patch changeset.

How did you test this change?

• On unchanged next@215c3652, 20 new cases fail and two controls pass. All 22 pass with the fix. Together with the existing netting regressions, all 31 focused tests pass in development, production and observe configurations.
• Signals JavaScript builds for all three tiers and source declarations pass (npm run build:js, npm run types in packages/signals).
• Full signals suite: 431 files / 6,192 tests pass, with 21 expected failures and two skips (vitest run --maxWorkers=2).
• Prettier checks, git diff --check and node scripts/check-release-invariants.mjs pass.
• Rolldown 1.2.11: +22 minified bytes in the core/createStore/until-refresh scenarios, +44 with isPending/latest. Brotli changes range from -7 to +13 bytes. All four signals size gates pass; no caps changed.
• Local production timings on Node 24.19.0 remain noisy. Eight fresh-process pairs per configuration gave median paired chain differences of -1.5% for native modules and +0.7% for the bundled core. Other bundled chain/diamond/netted measurements ranged from -4.9% to +5.3%. These measurements do not establish performance equivalence or application performance.

[changeset-bot]

🦋 Changeset detected

Latest commit: 6b2a19e

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

This PR includes changesets to release 15 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
board-example	Patch
canvas-example	Patch
server-dbmon-example	Patch
todos-server-example	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

[codspeed]

Merging this PR will regress 1 benchmark

1 warning

⚠️ 2 benchmarks are dominated by system calls

System calls cannot be simulated consistently and are left out of the measure. The reported value understates the real cost. Switch to the Walltime instrument to measure them.

Learn more about system calls

See affected benchmarks

⚡ 1 improved benchmark
❌ 1 regressed benchmark
✅ 198 untouched benchmarks
⏩ 5 skipped benchmarks¹

Warning

Please fix the performance issues or acknowledge them on CodSpeed.

Performance Changes

	Benchmark	BASE	HEAD	Efficiency
❌	spread store record × 500	31.1 ms	33.2 ms	-6.32%
⚡	build + consume	608.9 µs	287.9 µs	×2.1

Tip

Investigate this regression by commenting @codspeedbot fix this regression on this PR, or directly use the CodSpeed MCP with your agent.

---

_{Comparing GabbeV:fix/signed-zero-staged-read (6b2a19e) with next (215c365)}

Footnotes

1. 5 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩