fix(signals): a lazy memo read only through a rebuilt child memo is not released mid-pass (#4142) by ryansolid · Pull Request #4164 · solidjs/solid

ryansolid · 2026-10-19T22:26:45Z

Summary

Fixes #4142, the last open regression gating the next RC (since rc.14).

Before: a lazy async memo read on both sides of a conditional, as in {user() && <User data={user()} />} under <Show>/<Loading>, fetched forever and the boundary never left its fallback.

After: it fetches once and shows its content.

Cause. The compiler turns the condition into a child memo of the hole's pass, and that child memo is the lazy memo's only subscriber. While the fetch is pending the pass never commits, so the next pass tears the old frame's children down on the spot (recompute, CONFIG_STAGED) before it builds its own. That teardown unlinked the lazy memo's last subscriber, and unlinkSubs released it right away. The new condition memo's read then recomputed it from scratch, the fetch went pending again, and the cycle repeated.

How. During that one teardown, last-one-out releases are collected (deferredReleases in graph.ts) instead of run. When the pass ends, recompute releases each one that is still unobserved, reusing releaseIfSettledUnobserved. A lazy node the new pass reads again stays alive. One it stopped reading is still released at the end of the same pass. No public API change.

dormant-memo-chain-3554.test.ts: one expectation changes. In the "made dormant and reawakened while its owner's held children drain" case, lazy1's last reader dies in the same teardown, so lazy1 is no longer made dormant mid-drain and reawakened into the new frame by the cleanup's read. It now dies with the frame it belongs to. The test's point still holds: the drain runs to the end, sibling1 is cleaned, and sibling2 is cleaned on dispose.

Size: +104 B minified on the core floor, about +65 B brotli. Most of it is mangled-name churn from the two new module bindings: the code itself is about 80 B, and I tried three shapes before settling on this one. Hello world (app: compiled floor) stays under 10 KB brotli at 9.86 KB. Caps are raised to measured + 10 B with dated ledger notes, and each recorded minified size is set to the measured value.

Size-Exception: deferring lazy-memo releases until a pending pass ends stops the infinite re-fetch loop in #4142; the maintainer accepted the +104 B minified on 2026-10-19.

How did you test this change?

packages/signals/tests/lazy-memo-rebuilt-reader-4142.test.ts:
- The rebuilt-child-memo case: fails on next (11 runs, nothing shown), passes here.
- A guard that a lazy memo the rebuilt pass stops reading is still released at the end of that pass.
packages/web/test/lazy-memo-loading-loop-4142.spec.tsx: the report's playground, driven through the checkbox's change event. It loops on next and fetches once here.

Suites, run locally after turbo build:

Suite	Result
signals	5368 pass
web client	1392 pass
web hydrate	520 pass
solid	877 pass
web server	1732 pass; the only failure is the known Node 22 `server-functions-adapter-request` one

Generated by Claude Code

…ot released mid-pass (#4142) A pass whose previous pass never committed tears that frame's children down on the spot, before it builds its own. When a lazy memo's only subscriber was one of those children (the compiled condition memo of `user() && <User data={user()} />`), the teardown released it, the new child's read recomputed it from scratch, the async fetch went pending again, and the hole looped behind its Loading fallback forever. The teardown now collects those last-one-out releases and runs them when the pass ends, so a node the pass reads again stays alive and one it stopped reading is still released. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018LY7F9czoRqdMEvk2QwPr1

changeset-bot · 2026-10-19T22:26:50Z

🦋 Changeset detected

Latest commit: 0e272be

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

This PR includes changesets to release 14 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
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

github-actions · 2026-10-19T22:29:46Z

Size (brotli, eager entry chunk)

scenario	head	vs base	minified vs base	minified vs recorded	cap		lazy chunks (not counted)
signals: core floor (createSignal/Memo/Effect/Root/flush)	7.34 KB	+53 B (+0.7%)	+104 B	0 B	7.35 KB	✅
signals: + createStore	14.88 KB	+13 B (+0.1%)	+114 B	0 B	14.90 KB	✅
signals: + isPending/latest	10.10 KB	+51 B (+0.5%)	+106 B	0 B	10.12 KB	✅
signals: + until/refresh (truth readers, no optimism)	7.94 KB	+51 B (+0.6%)	+103 B	0 B	7.95 KB	✅
app: render + one signal (the simple-app floor)	9.64 KB	+34 B (+0.4%)	+106 B	0 B	9.66 KB	✅
app: hydrating (no stores) with Show/For/Loading/Errored/lazy	18.27 KB	+23 B (+0.1%)	+107 B	0 B	18.29 KB	✅	lazy-page.js 0.04 KB
app: hydrating + every store primitive family	30.29 KB	+55 B (+0.2%)	+114 B	0 B	30.30 KB	✅	lazy-page.js 0.04 KB
app: CSR with Show/For/Loading/Errored/lazy	13.23 KB	+7 B (+0.1%)	+109 B	0 B	13.24 KB	✅	lazy-page.js 0.04 KB
app: CSR, observe tier (same app on the `observe` artifacts)	14.97 KB	+45 B (+0.3%)	+108 B	0 B	14.99 KB	✅	lazy-page.js 0.04 KB
app: CSR, observe tier + attribution engine enabled	29.57 KB	+72 B (+0.2%)	+102 B	0 B	29.59 KB	✅	lazy-page.js 0.04 KB
app: compiled floor (one template, one text hole, one delegated click)	9.85 KB	+43 B (+0.4%)	+106 B	0 B	9.86 KB	✅
app: compiled CSR (JSX todo app: spread/merge/omit, events, class/style, keyed For, Show, Loading + lazy, store)	26.09 KB	+50 B (+0.2%)	+103 B	0 B	26.10 KB	✅	stats.js 0.18 KB
app: compiled hydrating (the same JSX todo app through hydrate(), compiled hydratable)	32.11 KB	+117 B (+0.4%)	+100 B	0 B	32.13 KB	✅	stats.js 0.20 KB
frames: eager client consumer (frames client + transport, lazy codec)	11.75 KB	0 B	0 B	0 B	11.76 KB	✅
page: base server components (hydrating + dynamic + frames + sf reference)	33.15 KB	+43 B (+0.1%)	+103 B	0 B	33.16 KB	✅	assets.js 0.78 KB, decode.js 6.25 KB, lazy-page.js 0.04 KB, regions.js 1.07 KB, trace.js 9.02 KB, wire.js 2.78 KB
page: live server components (base + live/GET + action + isPending/latest)	37.44 KB	+83 B (+0.2%)	+116 B	0 B	37.46 KB	✅	assets.js 0.78 KB, decode.js 6.25 KB, lazy-page.js 0.04 KB, regions.js 1.06 KB, trace.js 9.03 KB, wire.js 2.78 KB
page: compiled base server components (the base page as JSX: templates with class/style/attributes/events, For/Show; no spread)	35.17 KB	−2 B (−0.0%)	+110 B	0 B	35.19 KB	✅	assets.js 0.78 KB, decode.js 6.25 KB, regions.js 1.06 KB, sc-comments.js 0.20 KB, trace.js 9.04 KB, wire.js 2.77 KB
page: compiled live server components (the compiled base page + live/GET + action + isPending/latest)	41.29 KB	+49 B (+0.1%)	+111 B	0 B	41.31 KB	✅	eager (counted): web.js 21.91 KB; assets.js 0.78 KB, decode.js 6.25 KB, regions.js 1.06 KB, sc-comments.js 0.19 KB, trace.js 9.08 KB, wire.js 2.77 KB
page: base + router (base page + @solidjs/router: createRouter, two routes, preload, useNavigate)	40.70 KB	+86 B (+0.2%)	+104 B	0 B	40.71 KB	✅	assets.js 0.78 KB, decode.js 6.25 KB, lazy-page.js 0.04 KB, regions.js 1.06 KB, server.js 1.02 KB, serverForms.js 3.59 KB, trace.js 9.13 KB, wire.js 2.77 KB
page: live + router (live page + @solidjs/router: createRouter, two routes, preload, useNavigate)	46.84 KB	−22 B (−0.0%)	+107 B	0 B	46.86 KB	✅	eager (counted): client.js 27.03 KB; assets.js 0.78 KB, decode.js 6.25 KB, lazy-page.js 0.04 KB, regions.js 1.07 KB, server.js 1.02 KB, serverForms.js 3.30 KB, trace.js 9.06 KB, wire.js 2.80 KB
server: floor (getRequestEvent + isServer)	1.33 KB	0 B	0 B	0 B	1.34 KB	✅
server: renderToString (the server-render floor)	20.62 KB	0 B	0 B	−23 B	20.59 KB	⚠️ over by 34 B, 43 B minified headroom

⚠️ Over the brotli cap within the minified allowance (passes)

server: renderToString (the server-render floor): over brotli cap by 34 B; minified 72,131 B vs 72,154 B recorded with the cap (−23 B) — 43 B of the 20 B minified allowance left; +0 B minified over this PR's base

_{Bundled with Rolldown (what Vite ships), brotli q11, decimal KB. A scenario fails only when it is over its brotli cap and its minified size is more than 20 B over the minified recorded with the cap; over the cap within that allowance is brotli layout noise and passes with a warning. Caps and their recorded minified in scripts/size/scenarios.js; the floor and page caps in floor-caps.json are frozen (lower only, or Size-Exception: in the PR body). npm run ratchet lowers caps per RC; it never raises one (scripts/size/README.md).}

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018LY7F9czoRqdMEvk2QwPr1

coveralls · 2026-10-19T22:37:38Z

Coverage Report for CI Build 39339297559

Coverage remained the same at 77.555%

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:	1347
Covered Lines:	1103
Line Coverage:	81.89%
Relevant Branches:	1050
Covered Branches:	756
Branch Coverage:	72.0%
Branches in Coverage %:	Yes
Coverage Strength:	27.42 hits per line

💛 - Coveralls

ryansolid · 2026-10-19T22:38:14Z

The size gate fails because this fix adds about +104 B minified (~60 B brotli) to every scenario that includes the core, which is over the 20 B allowance. That growth is real, not noise. The caps stay as they are until the maintainer accepts the cost. Then I'll raise them from CI's numbers with ledger notes and add the Size-Exception: line.

— Claude via Claude Code

Generated by Claude Code

codspeed · 2026-10-19T22:44:54Z

Merging this PR will not alter performance

⚠️

1 benchmark spent significant time in system calls

System calls cannot be consistently instrumented, so they are not included in the measure, which understates the real cost. Please switch to the Walltime instrument to accurately measure system calls.

Measurement and system calls

readAllowed

✅ 196 untouched benchmarks
⏩ 5 skipped benchmarks¹

_{Comparing fix/lazy-memo-loop-4142 (0e272be) with next (a9b9f9e)}

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. ↩

ryansolid · 2026-10-19T22:55:45Z

CodSpeed flags mount 4000 rows › memo + sync render effect only (reference) at −10.6%. That benchmark is known to be noisy on this repo: it showed +13.1% on #4159, which doesn't touch this path, and CodSpeed warns that this comparison used a different base commit (667dcff instead of a9b9f9e) and a different runtime environment. Mounting never goes through the teardown path this PR changes, so I'm treating it as noise rather than a regression from this PR.

— Claude via Claude Code

Generated by Claude Code

Accepted by the maintainer on 2026-10-19: +104 B minified on the core floor. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018LY7F9czoRqdMEvk2QwPr1

Merge origin/next into fix/lazy-memo-loop-4142

3d529de

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018LY7F9czoRqdMEvk2QwPr1

chore(size): raise caps for the #4142 lazy-memo release fix

0e272be

Accepted by the maintainer on 2026-10-19: +104 B minified on the core floor. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018LY7F9czoRqdMEvk2QwPr1

ryansolid marked this pull request as ready for review October 19, 2026 23:04

ryansolid merged commit 6371b91 into next Oct 19, 2026
7 checks passed

This was referenced Oct 19, 2026

[2.0 rc.15] Reading lazy memo in both side of conditional expression causes infinite computation loop #4142

Closed

fix(signals): a removed row's cleanup can write during teardown (#4165) #4168

Merged

GabbeV mentioned this pull request Oct 20, 2026

fix(signals): restore deferred releases after cleanup throws #4176

Merged

ryansolid mentioned this pull request Oct 21, 2026

chore: #4183's changeset and re-record the compiled CSR size cap #4209

Merged

Navigation Menu

Repository navigation

fix(signals): a lazy memo read only through a rebuilt child memo is not released mid-pass (#4142) - #4164

fix(signals): a lazy memo read only through a rebuilt child memo is not released mid-pass (#4142)#4164
ryansolid merged 3 commits into
nextfrom
fix/lazy-memo-loop-4142

ryansolid commented Oct 19, 2026 •

edited

Loading

Uh oh!

changeset-bot Bot commented Oct 19, 2026 •

edited

Loading

Uh oh!

github-actions Bot commented Oct 19, 2026 •

edited

Loading

Uh oh!

coveralls commented Oct 19, 2026 •

edited

Loading

Uh oh!

ryansolid commented Oct 19, 2026

Uh oh!

codspeed Bot commented Oct 19, 2026 •

edited

Loading

Uh oh!

ryansolid commented Oct 19, 2026

Uh oh!

Uh oh!

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

Conversation

ryansolid commented Oct 19, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

How did you test this change?

Uh oh!

changeset-bot Bot commented Oct 19, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

🦋 Changeset detected

Uh oh!

github-actions Bot commented Oct 19, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Size (brotli, eager entry chunk)

⚠️ Over the brotli cap within the minified allowance (passes)

Uh oh!

coveralls commented Oct 19, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Coverage Report for CI Build 39339297559

Coverage remained the same at 77.555%

Details

Uncovered Changes

Coverage Regressions

Coverage Stats

💛 - Coveralls

Uh oh!

ryansolid commented Oct 19, 2026

Uh oh!

codspeed Bot commented Oct 19, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Merging this PR will not alter performance

Footnotes

Uh oh!

ryansolid commented Oct 19, 2026

Uh oh!

Uh oh!

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

ryansolid commented Oct 19, 2026 •

edited

Loading

changeset-bot Bot commented Oct 19, 2026 •

edited

Loading

github-actions Bot commented Oct 19, 2026 •

edited

Loading

coveralls commented Oct 19, 2026 •

edited

Loading

codspeed Bot commented Oct 19, 2026 •

edited

Loading