signals: a re-armed boundary owns content another hold staged (U-9)

[ryansolid]

Requested by Ryan · project thread

Summary

Before: a Loading boundary re-armed by on could not show its fallback when the content under it read a hold that another change made. The content joined that hold, so no fallback showed and the boundary reported pending. In the S-19 shape (a held tree under a re-armed boundary) the boundary redrew forever.

After: the re-armed boundary owns that content, as D-14 and D-16 rule. The fallback shows with the re-arming change, and the content reveals at the hold's commit, which no longer waits for it. A re-arm whose own change joins that hold still shows no fallback, because the frame waits on the same data (the "a read made pending before the flush joins its hold" pins are unchanged). LOADING_ON_OUTSIDE_HOLD is now reported only for a flight the change itself started.

This is U-9, the last of the re-arm deferrals. Nine pins flip from it.fails to it:

• the redraw loop;
• the two re-armed held mounts;
• the four "committed content under a re-armed boundary" shapes;
• §3b (two outside-reader shapes);
• §3c after the action (three shapes).

How:

• arm records what an earlier hold staged under the boundary (heldUnder, scheduler.ts) and flips. The boundary's output leaves that hold (unhold), so the fallback publishes now.
• A pass under a re-armed boundary that reads a hold is the boundary's, not the tick's. _fresh now answers for committed passes too (armedOver), at the held re-pass in recompute and in joinPass. A re-notified pending through such a pass does not join the flush, unless the change re-asked that flight.
• The seam folds a re-arm back into the hold when its own flush joined that hold: a fallback staged into the very hold its readers wait on is dropped.
• prune keeps the tree as a held reader while armed, which is the loop fix.

Size, measured locally against next @ a4b3391:

scenario	minified	brotli	brotli now / cap
signals: core floor	+0 B	−2 B	7,467 / 7,480
signals: + createStore	+0 B	−7 B	14,936 / 14,940
app: render + one signal (simple-app floor)	+0 B	+23 B	9,977 / 9,980
app: CSR with Show/For/Loading/Errored/lazy	+416 B	+140 B	13,181 / 13,070 (over)
app: hydrating (no stores)	+417 B	+187 B	18,074 / 17,930 (over)
app: hydrating + every store primitive family	+417 B	+94 B	29,631 / 29,570 (over)
app: CSR, observe tier	+416 B	+139 B	14,778 / 14,640 (over)
app: CSR, observe + attribution	+417 B	+138 B	29,370 / 29,240 (over)

The floors are held at 0 B minified. The core's hook calls (X && X(...) to X?.(...)) and removalStagedBy were golfed to pay for the two new hook arguments. The five boundary apps are over their caps by 61–144 B brotli. The caps are left untouched until the maintainer rules.

How did you test this change?

• pnpm --filter @solidjs/signals build:js && npx vitest run in packages/signals: 5229 passed, 2 expected fail (U-5, U-6), 2 skipped.
• Internal reader counts in four committed-content pins: held 1/1, flight 3/2. latest of held re-armed content is the staged value (c1 1, A11), as for a fresh boundary's content.
• Each piece was ablated against the full suite; the ones kept are the ones whose removal fails tests.
• node packages/signals/scripts/rules-index.mjs --check: current.
• node scripts/size/size.mjs (table above).

Size

Size-Exception: a re-armed boundary owns content another hold staged (D-14/D-16/U-9). The seven Loading-boundary app scenarios grow +416 to +418 B minified, so their caps are raised (hydrating with no stores is a floor-caps.json entry). The floors are 0 B minified. Accepted by the maintainer on 2026-10-14.

🤖 Generated with Claude Code

https://claude.ai/code/session_011JxYhsd8CeAgCp6h5DU8ic

---

Generated by Claude Code

---

Generated by Claude Code

[changeset-bot]

🦋 Changeset detected

Latest commit: af6186a

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

This PR includes changesets to release 13 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
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]

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)	7.47 KB	−2 B (−0.0%)	0 B	−8 B	7.48 KB	✅
signals: + createStore	14.94 KB	−7 B (−0.0%)	0 B	−22 B	14.94 KB	✅
signals: + isPending/latest	9.75 KB	−2 B (−0.0%)	0 B	−8 B	9.77 KB	✅
app: render + one signal (the simple-app floor)	9.98 KB	+23 B (+0.2%)	0 B	−8 B	9.98 KB	✅
app: hydrating (no stores) with Show/For/Loading/Errored/lazy	18.07 KB	+187 B (+1.0%)	+417 B	0 B	18.09 KB	✅	lazy-page.js 0.04 KB
app: hydrating + every store primitive family	29.63 KB	+94 B (+0.3%)	+417 B	0 B	29.65 KB	✅	lazy-page.js 0.04 KB
app: CSR with Show/For/Loading/Errored/lazy	13.18 KB	+140 B (+1.1%)	+416 B	0 B	13.20 KB	✅	lazy-page.js 0.04 KB
app: CSR, observe tier (same app on the observe artifacts)	14.78 KB	+139 B (+0.9%)	+416 B	0 B	14.79 KB	✅	lazy-page.js 0.04 KB
app: CSR, observe tier + attribution engine enabled	29.37 KB	+138 B (+0.5%)	+417 B	0 B	29.38 KB	✅	lazy-page.js 0.04 KB
app: compiled floor (one template, one text hole, one delegated click)	10.17 KB	+6 B (+0.1%)	0 B	−8 B	10.17 KB	✅
app: compiled CSR (JSX todo app: spread/merge/omit, events, class/style, keyed For, Show, Loading + lazy, store)	25.90 KB	+102 B (+0.4%)	+417 B	0 B	25.92 KB	✅	stats.js 0.18 KB
app: compiled hydrating (the same JSX todo app through hydrate(), compiled hydratable)	31.77 KB	+71 B (+0.2%)	+418 B	0 B	31.78 KB	✅	stats.js 0.20 KB
frames: eager client consumer (frames client + transport, lazy codec)	11.49 KB	0 B	0 B	0 B	11.49 KB	✅
page: base server components (hydrating + dynamic + frames + sf reference)	34.35 KB	+158 B (+0.5%)	+417 B	−985 B	34.33 KB	⚠️ over by 14 B, 1005 B minified headroom	assets.js 0.78 KB, bind.js 1.83 KB, decode.js 6.23 KB, lazy-page.js 0.04 KB, regions.js 0.80 KB, trace.js 9.00 KB, wire.js 0.93 KB
page: live server components (base + live/GET + action + isPending/latest)	38.11 KB	+145 B (+0.4%)	+417 B	−1054 B	38.15 KB	✅	assets.js 0.78 KB, bind.js 1.83 KB, decode.js 6.23 KB, lazy-page.js 0.04 KB, regions.js 0.80 KB, trace.js 9.01 KB, wire.js 0.93 KB
page: compiled base server components (the base page as JSX: templates with class/style/attributes/events, For/Show; no spread)	35.20 KB	+204 B (+0.6%)	+417 B	−2075 B	35.66 KB	✅	assets.js 0.78 KB, bind.js 1.83 KB, decode.js 6.23 KB, regions.js 0.80 KB, sc-comments.js 0.20 KB, trace.js 8.98 KB, wire.js 0.93 KB
page: compiled live server components (the compiled base page + live/GET + action + isPending/latest)	40.81 KB	+126 B (+0.3%)	+417 B	−2068 B	41.21 KB	✅	eager (counted): web.js 21.68 KB; assets.js 0.78 KB, bind.js 1.84 KB, decode.js 6.23 KB, regions.js 0.79 KB, sc-comments.js 0.19 KB, trace.js 8.97 KB, wire.js 0.93 KB
page: base + router (base page + @solidjs/router: createRouter, two routes, preload, useNavigate)	41.67 KB	+71 B (+0.2%)	+417 B	−991 B	41.79 KB	✅	assets.js 0.78 KB, bind.js 1.85 KB, decode.js 6.23 KB, lazy-page.js 0.04 KB, regions.js 0.80 KB, server.js 1.02 KB, serverForms.js 3.58 KB, trace.js 9.01 KB, wire.js 0.94 KB
page: live + router (live page + @solidjs/router: createRouter, two routes, preload, useNavigate)	47.47 KB	+142 B (+0.3%)	+411 B	−991 B	47.49 KB	✅	eager (counted): client.js 27.89 KB; assets.js 0.78 KB, bind.js 1.84 KB, decode.js 6.23 KB, lazy-page.js 0.04 KB, regions.js 0.81 KB, server.js 1.02 KB, serverForms.js 3.30 KB, trace.js 8.97 KB, wire.js 0.94 KB
server: floor (getRequestEvent + isServer)	1.33 KB	0 B	0 B	0 B	1.34 KB	✅
server: renderToString (the server-render floor)	20.53 KB	0 B	0 B	0 B	20.55 KB	✅

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

• page: base server components (hydrating + dynamic + frames + sf reference): over brotli cap by 14 B; minified 105,719 B vs 106,704 B recorded with the cap (−985 B) — 1005 B of the 20 B minified allowance left; +417 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).}

[ryansolid]

The check job fails at the size gate (node gate.mjs size-head.json size-base.json). This PR's own bytes cause it: seven Loading-boundary app scenarios go 61–144 B brotli over their caps (CSR, hydrating, hydrating+stores, observe, observe+attribution, compiled CSR, compiled hydrating). The hello-world floors stay at 0 B minified.

I haven't fixed this because the choice is a maintainer's call: raise those caps, golf further, or drop U-9 and file it as an issue. I've asked in the project thread and am leaving the caps alone until that's answered. The PR stays a draft and watched.

---

Generated by Claude Code

[coveralls]

Coverage Report for CI Build 38588969410

Coverage remained the same at 77.515%

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:	1326
Covered Lines:	1084
Line Coverage:	81.75%
Relevant Branches:	1040
Covered Branches:	750
Branch Coverage:	72.12%
Branches in Coverage %:	Yes
Coverage Strength:	27.49 hits per line

---

💛 - Coveralls

[codspeed]

Merging this PR will not alter performance

⚠️ 2 benchmarks 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
• propagation:diamond

✅ 195 untouched benchmarks

---

_{Comparing claude/u9-rearm-boundary-scope (af6186a) with next (f048691)}

[ryansolid]

CodSpeed's first run flagged one benchmark: mount-rows "memo + sync render effect only (reference)", down 13.6%, with a runtime-environment warning. The only hot-path change here sits inside the held re-pass branch in recompute, and mount-rows never enters it.

Update: I re-ran the same commit (c2513e3) and all 195 benchmarks came back untouched, so the first reading was runner noise.

---

Generated by Claude Code