[brenelz]

Summary

A delegated event keeps an unmounted component alive in Chrome for the life of the page.

eventHandler simulated currentTarget with an accessor defined on each event, and each event got a new getter closure. V8 keeps the first such getter in the descriptor array that every event of that class shares. A heap snapshot of a page after one keydown shows the getter on the KeyboardEvent map, and its closure holds the event, the event's target and the nodes of the walk. #4325 deletes the override when the walk ends, but the delete does not remove the getter from the shared shape, so rc.21 still leaks.

currentTarget is now a configurable data property. It is defined right before each delegated handler runs, through a define(name, value) helper that replaces retarget, so no getter is created. The deletes from #4325 stay, and native listeners that run later in the dispatch still read the native values. Each client scenario is 42 B smaller minified.

I found this through vercel-labs/ui, which carries a pnpm patch on @solidjs/web that shares one getter across events. Its Solid leak spec mounts and unmounts each component 50 to 200 times and counts detached DOM nodes and heap growth. On rc.21 without that patch, 4 of its 47 tests fail on a built playground: combobox, slider, resizable, and tabs after an arrow key.

How did you test this change?

The new test in packages/web/test/delegated-event-restore.spec.tsx reads the event's own currentTarget descriptor inside two nested delegated handlers. It expects no getter and the handler's own node as the value. It fails on next, where the descriptor has a getter.

After pnpm build, in packages/web:

• npx vitest run: 266 files, 1,900 tests pass.
• npx vitest run --config vite.config.hydrate.mjs: 147 files, 779 tests pass.
• npx vitest run --config vite.config.server.mjs: 228 files pass. test/server/foreign-raw-text-4704.spec.tsx does not compile with my local compiler binary, which is older than the fix for <style> and <script> children inside MathML are not escaped in SSR (and static DOM templates), allowing markup injection #4704.

Size, from node size.mjs --no-gate --json on this branch and on its base, then node gate.mjs head.json base.json: every scenario that includes the client runtime is 42 B smaller minified. "page: compiled live server components" is over its cap on the base as well (45,045 B against 44,950 B with my local compiler binary) and stays over.

In vercel-labs/ui on rc.21 with its patch removed, I applied this change to the installed @solidjs/web builds and ran npx playwright test specs/leaks.spec.ts --project solid --workers 1 --retries 0 against its dev server with performance tracks off: 47 of 47 pass. I have not rerun the built playground with the change applied.

🤖 Generated with Claude Code

[changeset-bot]

🦋 Changeset detected

Latest commit: 6ec96cf

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

This PR includes changesets to release 15 packages

Name	Type
@solidjs/web	Patch
@solidjs/babel-plugin	Patch
@solidjs/diagnostics	Patch
@solidjs/element	Patch
@solidjs/h	Patch
@solidjs/html	Patch
test-integration	Patch
board-example	Patch
canvas-example	Patch
server-dbmon-example	Patch
todos-server-example	Patch
@solidjs/compiler	Patch
@solidjs/signals	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

[codspeed]

Merging this PR will degrade performance by 11.91%

3 warnings

⚠️ Upgrade your CodSpeed plugin for Node.js 24

@codspeed/* 5.4.0 is only tested on Node.js 18, 20, and 22, so your Node.js 24 measurements may be inaccurate. Upgrade your @codspeed/* packages to 6.0.0 or later.

See affected benchmarks

⚠️ 3 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 benchmark with a significant change ran on different hardware

Its base and head results were measured on machines with a different CPU, core count or memory, so the change might not come from the code. Run base and head on the same runner type.

Learn more about Macro Runners

See affected benchmarks

❌ 1 regressed benchmark
✅ 200 untouched benchmarks
⏩ 5 skipped benchmarks¹

Warning

Please fix the performance issues or acknowledge them on CodSpeed.

Performance Changes

	Benchmark	BASE	HEAD	Efficiency
❌	projection derive: write one NESTED field (reference)	2.2 ms	2.5 ms	-11.91%

Tip

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

---

_{Comparing brenelz:fix/web-shared-currenttarget (6ec96cf) with next (f00ce6f)}

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. ↩