rc.14: an effect that catches NotReadyError from an async store is never re-run when the store settles
维护者通常 1 天内回复
还没有人认领这个 Issue。
评估
- 难度
- 4/5
- 预计耗时
- 3-5 天
- 新手友好度
- 25/100
- Issue 类型
- 缺陷
- 描述清晰度
- 描述清楚
- 活跃度
- 活跃
- 技术栈
- javascript, typescript
- 领域
- frontend
调研方向
Start with the headless repro.mjs in the issue, which pins @solidjs/signals to 2.0.0-rc.13 and then rc.15 to compare effect re-runs. The payload names no source files, so locate where NotReadyError is thrown for async store derives and where an effect is scheduled to re-run when that derive settles. Done means the effect with try/catch records ["not-ready","len=1"] again on the current version, matching rc.13, and the without-try/catch case still passes.
由索引模型根据 Issue 内容生成。
描述
Summary
@solidjs/[email protected] and rc.15 (dev and prod builds). Regression since rc.13.
An effect whose compute reads an async-seeded store (createStore(async fn, [])) and catches the NotReadyError that read throws runs once with its fallback value and is never re-run when the store's seed settles. The store itself settles: an untracked read afterwards returns the data. The same effect without the try/catch holds and then runs once with the settled value, as before. An async createMemo read the same way is still re-run on rc.15, so this looks specific to async store derives.
const [rows] = createStore(async () => await request, []);
createEffect(
() => { try { return `len=${rows.length}`; } catch { return "not-ready"; } },
(v) => seen.push(v),
);
// rc.13: ["not-ready", "len=1"] rc.14 / rc.15: ["not-ready"], forever
Reproduction (headless, only @solidjs/signals)
mkdir solid-catch-notready && cd solid-catch-notready
npm init -y >/dev/null && npm pkg set type=module && npm i @solidjs/[email protected] >/dev/null
cat > repro.mjs <<'JS'
// An effect whose compute catches NotReadyError from an async-seeded store
// is re-run when the seed settles (rc.13) — or never again (rc.14, rc.15).
import { createEffect, createRoot, createStore, flush } from "@solidjs/signals";
const tick = (ms) => new Promise((r) => setTimeout(r, ms));
for (const caught of [false, true]) {
let resolve;
const request = new Promise((r) => (resolve = r));
const seen = [];
const dispose = createRoot((dispose) => {
const [rows] = createStore(async () => await request, []);
createEffect(
() => {
if (!caught) return `len=${rows.length}`;
try { return `len=${rows.length}`; } catch { return "not-ready"; }
},
(v) => { seen.push(v); },
);
return dispose;
});
flush();
resolve([{ id: "a" }]);
await tick(10);
flush();
console.log(`${caught ? "with try/catch " : "without try/catch"}: ${JSON.stringify(seen)}`);
dispose();
}
JS
node repro.mjs
@solidjs/signals |
without try/catch | with try/catch |
|---|---|---|
| 2.0.0-rc.13 | ["len=1"] |
["not-ready","len=1"] |
| 2.0.0-rc.14 | ["len=1"] |
["not-ready"] |
| 2.0.0-rc.15 | ["len=1"] |
["not-ready"] |
Replacing the store with const memo = createMemo(async () => await request) (reading memo().length) gives ["not-ready","len=1"] on rc.15.
(Pin @solidjs/signals directly: npm i [email protected] alone installs the newest signals through its caret dependency.)
Where we hit it
@dschz/solid-flow pins its async store factories (createNodeStore(async () => …), a thin wrapper over createStore(fn, [])) with this exact effect, which records "not-ready" until the first value and then the row count. Two of its tests fail on rc.14 and rc.15. The library itself never catches NotReadyError, so for us this is a test idiom, but a user's effect that degrades gracefully while data loads would stay on its fallback forever.
If catching NotReadyError in a compute is not a supported pattern, a dev diagnostic for it would help: today the effect silently never updates.
- 主要语言
- TypeScript
- 星标
- 36.2k
- 派生
- 1.1k
- 平均合并
- 10 小时 34 分钟
- 30 天内合并 PR
- 364
环境准备
- 没有 Dockerfile 或 Docker Compose 文件
- 有 Pull Request 模板
- 阅读贡献指南
从这里开始
- 先读完整个 Issue,再读项目的贡献指南。
- 在 Issue 下留言说明你要接手 —— 这能避免两个人做同样的事。
- Fork 仓库,在一个分支上完成修改。
- 提交 Pull Request,并在描述里引用这个 Issue 编号。
solidjs/solid 的其他 Issue
-
`lazy()`: remounting an already-loaded component renders empty (or the Suspense fallback) since 1.9.14可能已有人在做 @itisrmk 于 5 天前认领。 未关闭1.x
难度 2/5 1-3 小时 新手友好度 62/100
维护者通常 1 天内回复
-
难度 4/5 3-5 天 新手友好度 44/100
维护者通常 1 天内回复
-
enhancement
难度 5/5 一周以上 新手友好度 35/100
维护者通常 1 天内回复
-
`createRoot`'s dispose function keeps `fn` alive, so `<For>`/`<Index>` rows retain removed list items可能已有人在做 @boombang 于 7 天前认领。 未关闭1.x
难度 4/5 1-2 天 新手友好度 45/100
维护者通常 1 天内回复
-
enhancement
难度 5/5 一周以上 新手友好度 35/100
维护者通常 1 天内回复
相似的 Issue
-
数据界面的横向滚动条缺失未关闭
难度 2/5 1-3 小时 新手友好度 62/100
zilliztech/attu#1037 ·
-
clawsweeper:fix-shape-clear clawsweeper:queueable-fix clawsweeper:source-repro impact:ux-friction issue-rating: 🦞 diamond lobster no-stale P3
难度 2/5 1-3 小时 新手友好度 76/100
openclaw/openclaw#181212 · 1 条评论 · 1 个 reaction ·
维护者通常 1 天内回复
-
难度 2/5 1-3 小时 新手友好度 78/100
apify/apify-mcp-server#1684 ·
维护者通常 1 天内回复
-
难度 2/5 1-3 小时 新手友好度 82/100
Automattic/data-liberation-agent#953 ·
维护者通常 1 天内回复
-
难度 2/5 1-3 小时 新手友好度 78/100
siteboon/claudecodeui#1642 ·
维护者通常 5 天内回复