owner stringclasses 8
values | repo stringclasses 9
values | number int32 1 155k | node_id stringlengths 18 32 | is_pull_request bool 2
classes | title stringlengths 1 754 | body stringlengths 0 259k | state stringclasses 5
values | state_reason stringclasses 7
values | author stringlengths 0 39 | created_at timestamp[us]date 2000-06-06 02:40:44 2026-04-01 17:49:29 | updated_at timestamp[us]date 2010-06-29 02:24:09 2026-04-01 17:52:48 | closed_at timestamp[us]date 2000-06-07 02:36:03 2026-04-01 16:55:56 ⌀ | labels unknown | assignees unknown | milestone_title stringclasses 812
values | milestone_number int32 0 434 | reactions unknown | comment_count int32 0 2.47k | locked bool 2
classes | lock_reason stringclasses 5
values | detail_fetched_at timestamp[us]date 2026-03-29 04:12:38 2026-04-01 17:51:16 ⌀ |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
facebook | react | 35,300 | I_kwDOAJy2Ks7ci2BB | false | Bug : Component not rendering | Issue while rendering components. | closed | completed | mausam-005 | 2025-12-05T19:30:06 | 2025-12-09T16:05:07 | 2025-12-09T16:05:07 | null | null | 0 | {
"+1": 2,
"-1": 1,
"hooray": 1,
"rocket": 1
} | 1 | false | 2026-03-29T04:51:59.360000 | ||
facebook | react | 34,382 | I_kwDOAJy2Ks7Jtr4i | false | Bug: jsdom issues with React 19.1.1 | When using testing-library / jsdom, the special characters created by getId break nwsapi, which therefore breaks usages of `screen.findByRole`. Fortunately, #33422 already fixes this issue, but is unreleased so far as I can tell. Please release it soon, thank you. | closed | not_planned | jcary741 | 2025-09-04T15:07:22 | 2025-12-11T06:09:19 | 2025-12-11T06:09:19 | [
"Status: Unconfirmed",
"Resolution: Stale"
] | null | 0 | {
"heart": 3
} | 2 | false | 2026-03-29T04:51:59.447789 | ||
facebook | react | 34,353 | I_kwDOAJy2Ks7I4ksB | false | Bug: | ⨯ [Error: Expected a suspended thenable. This is a bug in React. Please file an issue.] {
digest: '1013962398'
}
React version: "react": "19.1.0","react-dom": "19.1.0","next": "15.4.4",
failed to pipe response – This usually happens in server-side rendering (SSR) with frameworks like Next.js. React tried to stream HTML to the client but ran into a problem with a component that should suspend (like a Suspense or use async call).
Expected a suspended thenable – React was expecting a Promise-like object to suspend rendering, but it got something else. Essentially, a component is not returning what React expects for a suspended state.
| closed | not_planned | mohamedAlmosanid | 2025-08-31T13:41:14 | 2025-12-08T18:08:28 | 2025-12-08T18:08:28 | [
"Status: Unconfirmed",
"Resolution: Stale"
] | null | 0 | {} | 2 | false | 2026-03-29T04:51:59.528842 | ||
facebook | react | 11,171 | MDU6SXNzdWUyNjQxNDU3NTY= | false | Should React use requestAnimationFrame by default? | Consider the following sample code: (pasted [here too](https://gist.github.com/dakom/e68503e59c7a73c3bc926b7e8ae65a07))
```
class extends React.Component {
private canRender: boolean = false;
private latestData: any;
constructor(props) {
super(props);
let nJobs = 0;
let lastRenderTime: number;
props.someObservableThing.listen(data => {
nJobs++;
this.latestData = data;
if (this.canRender) {
const now = performance.now();
this.canRender = false;
this.setState({
data: this.latestData,
jobsPerRender: nJobs,
fps: (lastRenderTime === undefined) ? 0 : 1000 / (now - lastRenderTime)
});
nJobs = 0;
lastRenderTime = now;
}
});
this.state = {};
}
/* Lifecycle */
componentDidMount() {
this.canRender = true;
}
componentDidUpdate() {
this.canRender = true;
}
render() {
outputStats(this.state);
return this.state.data === undefined ? null : <View {...this.state.data} />
}
}
```
When outputStats is hit - I'm getting framerates of like 2000fps. In other words `requestAnimationFrame` does _not_ seem to be a limiter for react itself.
Is this correct?
(as a slightly separate topic- if that is true, for animation things do you think it would be good to simply wrap the `if (this.canRender) {}` block in a `requestAnimationFrame()`? I guess that's not really a React question though since the observableThing could also be capped via ticks...)
| closed | completed | dakom | 2017-10-10T08:47:25 | 2025-12-10T22:00:56 | 2018-08-30T14:02:33 | [
"Component: DOM",
"Type: Discussion",
"Component: Reconciler"
] | null | 0 | {
"+1": 1
} | 36 | false | 2026-03-29T04:51:59.573964 | ||
facebook | react | 31,932 | I_kwDOAJy2Ks6kouXe | false | [Compiler]: React Compiler should error on document.getElementById call during render | ### What kind of issue is this?
- [ ] React Compiler core (the JS output is incorrect, or your app works incorrectly after optimization)
- [ ] babel-plugin-react-compiler (build issue installing or using the Babel plugin)
- [ ] eslint-plugin-react-compiler (build issue installing or using the eslint plugin)
- [ ] react-compiler-healthcheck (build issue installing or using the healthcheck script)
### Link to repro
https://playground.react.dev/#N4Igzg9grgTgxgUxALhAMygOzgFwJYSYAEAIngIYA2EA5gOp44AWAChDDlQBTBF5gB5AA4JiAXwCURYAB1iROITA4FhTnkwIYRALxEcATxEQ0RACYQ4UALaiVAQh16ZILGYRoNCMy6IB+c0sbOwA6GgQcAFFKBFtMHAAhAwBJMy4AciF2TkoAWhgICBx0qWQiTChKSgBuOSIiGAjYYi46+qIAHjIqWiIIEUwdYH5hUTEAPjb2zu7qGhC2DipVePIvGCHFVfWJqemZijmQgGE1O3GACQQqiA6AelnaE7P4yflp+8f5xZy3j4fDrQ-kQJLVMGIQGIgA
### Repro steps
**Steps to reproduce**
1. Create a simple dialog component that uses Portal with a dynamic container:
```jsx
function DialogWithPortal({ isOpen }) {
const container = typeof document !== "undefined" ? document.getElementById('portal-root') : null;
return (
<Dialog open={isOpen}>
<Dialog.Portal container={container}>
<Dialog.Content>Hello</Dialog.Content>
</Dialog.Portal>
</Dialog>
);
}
```
2. Mount the component when portal container exists in DOM:
```jsx
// DOM has: <div id="portal-root"></div>
<DialogWithPortal isOpen={true} />
```
3. Observe that dialog doesn't open because compiler memoizes initial null container value
**Expected behavior:**
- document.getElementById() should be evaluated on each render to find the portal container
**Actual behaviour:**
- Compiler memoizes the initial document.getElementById() call
- If container doesn't exist on first render, null is cached permanently
- Subsequent renders use cached null value even after container is added to DOM
**The issue is in how the compiler transforms the code:**
```jsx
// Initial DOM query is memoized and never rechecked
let t1;
if ($[0] === Symbol.for("react.memo_cache_sentinel")) {
t1 =
typeof document !== "undefined"
? document.getElementById("portal-root")
: null;
$[0] = t1;
} else {
t1 = $[0];
}
const container = t1;
```
### How often does this bug happen?
Every time
### What version of React are you using?
^18.2.0
### What version of React Compiler are you using?
19.0.0-beta-201e55d-20241215 | closed | not_planned | rishitells | 2024-12-29T07:20:54 | 2025-12-11T16:08:29 | 2025-12-11T16:08:29 | [
"Component: React Compiler",
"Resolution: Stale",
"Compiler: todo",
"Compiler: New Validation Ideas"
] | null | 0 | {} | 8 | false | 2026-03-29T04:51:59.647216 | ||
facebook | react | 34,369 | I_kwDOAJy2Ks7JcywS | false | [Compiler]: Invocation of hook at specific position breaks memoization | ### What kind of issue is this?
- [x] React Compiler core (the JS output is incorrect, or your app works incorrectly after optimization)
- [ ] babel-plugin-react-compiler (build issue installing or using the Babel plugin)
- [ ] eslint-plugin-react-compiler (build issue installing or using the eslint plugin)
- [ ] react-compiler-healthcheck (build issue installing or using the healthcheck script)
### Link to repro
https://playground.react.dev/#N4Igzg9grgTgxgUxALhAMygOzgFwJYSYAEAggBQAOMEFYAlEcADrFFyFg5FRgIyYBDALYIwRALxEqNMADoefOUIEUyCmBIB83XjHm7BIugG4WLIkQD0logElMANwhwB+QkQhoiOABZ4xPhAQANZErt5+YhQQYHhuxC4KYuqGokQiQhB4AF6uBMQ4EEQARjAIAsGy5lY2tl4AntBsEEIimFxxHlBcEBqZDghEnQLFEAM6fKlivSUIADYQAO5EYIFQcwAmAMqBiwA01dYRCMQpwmn+6QiZOQgbYWIIAB4UCLh3VawKWy0IvniYADmZDoZlY7EwnBWa02OyWEhWvwAYlhcPk1AZzvRTKxqmUcLBiKtoLDdkQAPxEAA8ACEJvwseJgGcRGAAL5WbTIIiYdZzHFslggPYgCFoPCAlAgEBsoA
### Repro steps
Invocation of hook at a certain position in the render function breaks memoization:
```javascript
function A(props) {
const usernames = props.users.map(user => user.username);
// Invocation of this hook at this position causes usernames memoization to break.
// If you comment it out or move it above usernames or below shouldShow,
// then usernames is memoized as expected.
useSomething()
const shouldShow = someFunction(usernames);
return shouldShow ? <B usernames={usernames} /> : null;
}
```
### How often does this bug happen?
Every time
### What version of React are you using?
19.0.0
### What version of React Compiler are you using?
19.1.0-rc.3 | closed | not_planned | cknitt | 2025-09-03T12:58:39 | 2025-12-11T06:09:20 | 2025-12-11T06:09:20 | [
"Component: React Compiler",
"Resolution: Stale"
] | null | 0 | {
"+1": 2
} | 6 | false | 2026-03-29T04:51:59.687345 | ||
facebook | react | 34,398 | I_kwDOAJy2Ks7J7QWD | false | Bug: Potential React 19.1.1 regression bug (ISO date strings in nested objects) |
React 19.1.1 has a potential regression bug where ISO date strings in object arrays cause ref handling errors during prop processing, manifesting as "Cannot read properties of undefined (reading 'ref')" errors in third-party libraries.
When passing props containing arrays of objects with ISO date strings to React components, React 19.1.1's internal prop processing corrupts its ref handling mechanism. This causes downstream ref access errors in third-party libraries like @restart/ui (used by react-bootstrap).
This was found in a Next.js application (Next 15.3.4) and the bug occurred in deployed/built environments but not in development environments.
React version: 19.1.1
## Steps To Reproduce
1. Create a React component that receives props containing an array of objects with ISO date strings
2. Upgrade to React 19.1.1
3. Render the component with data containing ISO dates
4. Observe ref handling errors
```
// This data structure causes the bug:
const problematicData = {
dataArray: [
{
id: 1,
date: "2024-08-27T00:00:00.000+0000", // ISO date string
},
{
id: 2,
date: "2024-09-26T00:00:00.000+0000", // ISO date string
}
// ... more entries with ISO dates
]
};
// Component that fails:
function MyComponent({ data }) {
return <div>{/* any content */}</div>;
// This triggers the bug:
<MyComponent data={problematicData} />
```
## The current behavior
Components with props passed that contain certain kinds of ISO date codes fail.
## The expected behavior
All props containing nested objects with ISO date codes regardless of timestamp are able to render without throwing ref errors.
| closed | not_planned | TrentHurlbut | 2025-09-05T14:25:14 | 2025-12-11T21:06:09 | 2025-12-11T21:06:09 | [
"Status: Unconfirmed",
"Resolution: Stale"
] | null | 0 | {
"heart": 1
} | 6 | false | 2026-03-29T04:51:59.764788 | ||
facebook | react | 31,744 | I_kwDOAJy2Ks6jFKFD | false | Bug: initial props will leak if a component renders nested elements, when an outer element never rerenders. | JSX code like this will leak that `something` value:
```ts
<div>
<SubComponent foo={something} />
</div>
```
As the component and the sub component re-render. That div will never change, and will hold on to that initial prop via `div.__reactProps$[SECRET].children[0].props.foo`.
This is not only unexpected. In cases it can be very difficult to work around.
Here at Framer (http://framer.com) we use a tree like structure that represents the user document. Which potentially can be huge, like 1G or more. Sometimes this has to be reloaded from the server. At that point we cannot afford having the old document lingering around. But it is react that will continue to retain it due to this issue.
React version: 18.2.38 (But has been in react since at least 17.)
See also: https://github.com/facebook/react/issues/25658 (now closed as stale) and other reports filed in the past. I hope that by filing this one, with such a small reproduction, helps someone to dig in and hopefully fix it? 🙏
## Steps To Reproduce
Click the "next" button in the code example. That switches from a huge array to a large one. But the memory use won't drop.
Link to code example: https://codesandbox.io/p/sandbox/leak-static-element-children-props-forked-zdm83w
Or the simplified version:
```ts
import React from "react";
function SubComponent({ prop }: { prop: number[] }) {
return <p>array length: {prop.length}</p>;
}
function Test({ prop }: { prop: number[] }) {
return (
<div className="holder">
<SubComponent prop={prop} />
</div>
);
}
export default function App() {
// Initially, a very large state.
const [state, setState] = React.useState(
Array.from({ length: 20_000_000 }, () => Math.random())
);
return (
<>
<Test prop={state} />
<button
onClick={() => {
// Any next state is a lot smaller.
setState(Array.from({ length: 1_000_000 }, () => Math.random()));
}}
>
Next
</button>
</>
);
}
```
## The current behavior
Leaks memory:
<img width="935" alt="memory-leak" src="https://github.com/user-attachments/assets/8b1339c5-9a23-48d5-b007-559a856f8052" />
## The expected behavior
React internals should not be retaining properties that the code using the library has moved away from. | closed | not_planned | onnlucky | 2024-12-12T14:14:36 | 2025-12-11T13:17:40 | 2025-12-11T13:17:40 | [
"Status: Unconfirmed",
"Resolution: Stale"
] | null | 0 | {
"+1": 10
} | 7 | false | 2026-03-29T04:51:59.834464 | ||
facebook | react | 12,615 | MDU6SXNzdWUzMTQzOTQ0MDQ= | false | Unexpected warning when hydrating with portal and SSR | **Do you want to request a *feature* or report a *bug*?**
*bug*
**What is the current behavior?**
Given the following (simplified) snippet:
```jsx
class HoverMenu extends React.Component {
render() {
if (typeof document === 'undefined') return null
const root = document.getElementById('root')
return ReactDOM.createPortal(<div>Hello World</div>, root)
}
}
class Para extends React.Component {
render() {
return (
<span>
Some Text
<HoverMenu />
</span>
)
}
}
```
where `div#root` is a valid `div` that exists, the following error is shown when hydrating after SSR:
`Warning: Expected server HTML to contain a matching <div> in <span>`
The warning goes away if I update the definition of `HoverMenu` to:
```jsx
class HoverMenu extends React.Component {
componentDidMount() {
this.setState({ isActive: true })
}
render() {
const { isActive} = this.state
if (!isActive) return null
const root = document.getElementById('root')
return ReactDOM.createPortal(<div>Hello World</div>, root)
}
}
```
I'd prefer not to do that because of the double rendering caused by `setState` in `componentDidMount`.
I don't quite understand what that error is telling me. No `<div />` is rendered server-side in either case. The error is particularly confusing, as the `HoverMenu` DOM `div` is not even rendered inside a DOM `span`. (I wonder if this is happening because `HoverMenu` is nested inside a React `span`.)
**What is the expected behavior?**
No error is thrown. Or, at least that the error message is clearer.
**Which versions of React, and which browser / OS are affected by this issue? Did this work in previous versions of React?**
Chrome 65
React 16.2
(SSR through Next 5.1)
| open | majelbstoat | 2018-04-15T07:50:53 | 2025-12-11T02:39:05 | null | [
"Type: Bug",
"Difficulty: medium"
] | null | 0 | {
"+1": 12
} | 29 | false | 2026-03-29T04:51:59.886082 | |||
facebook | react | 34,515 | I_kwDOAJy2Ks7MMyU1 | false | I can't use React DevTools | I recently installed React Developer Tools and I could not use it. I found out about it in my JavaScript Console and went ahead to install from this link by [React](https://react.dev/learn/react-developer-tools). Since I was using Google Chrome on Microsoft Windows 11 Home Single Language [Version 24H2] (OS Build 26100,4946) {Insider Version} I went and clicked on the button saying Install for Chrome ``` <a href="https://chrome.google.com/webstore/detail/react-developer-tools/fmkadmapgofadopljbjfkapdkoienihi?hl=en" target="_blank" rel="nofollow noopener noreferrer" class="inline text-link dark:text-link-dark border-b border-link border-opacity-0 hover:border-opacity-100 duration-100 ease-in transition leading-normal">Install for <strong class="font-bold">Chrome</strong></a> ``` and installed it, from [here](https://chromewebstore.google.com/detail/react-developer-tools/fmkadmapgofadopljbjfkapdkoienihi?hl=en) . Later on, when I tested on my own app that I was building. It did not work and showed a error saying [This page doesn’t appear to be using React.
If this seems wrong, follow the [troubleshooting instructions](https://github.com/facebook/react/tree/main/packages/react-devtools#the-react-tab-doesnt-show-up).] Directing me here still it did not work. And, For Your Information I was using http://localhost:5173
Thank You
Reshabh Ohall
(12 y/o, India)
<img width="1919" height="1079" alt="Image" src="https://github.com/user-attachments/assets/bf604acd-0a44-480e-92a3-384dd52bf0e7" />
<img width="1919" height="1079" alt="Image" src="https://github.com/user-attachments/assets/5dbc95a2-daa8-4638-bfb6-6220c1634ce2" />
<img width="1918" height="1079" alt="Image" src="https://github.com/user-attachments/assets/01b6bdec-3353-457f-b4cc-8620e4f27498" /> | closed | completed | Reshabh-Ohall | 2025-09-17T11:34:55 | 2025-12-12T04:39:37 | 2025-12-12T04:39:36 | null | null | 0 | {
"heart": 2
} | 3 | false | 2026-03-29T04:51:59.954862 | ||
facebook | react | 35,350 | I_kwDOAJy2Ks7dzDZg | false | [Compiler Bug]: 'use no memo' does not apply recursively | ### What kind of issue is this?
- [x] React Compiler core (the JS output is incorrect, or your app works incorrectly after optimization)
- [ ] babel-plugin-react-compiler (build issue installing or using the Babel plugin)
- [x] eslint-plugin-react-hooks (build issue installing or using the eslint plugin)
- [ ] react-compiler-healthcheck (build issue installing or using the healthcheck script)
### Link to repro
https://playground.react.dev/#N4Igzg9grgTgxgUxALhAgHgBwjALgAgBMEAzAQygBsCSoA7OXASwjvwGEJKoBbOgGTIAjBJTAAKYPhIQI+AL4BKfMAA6bfAHIoYBPjpyeCHhE0BudfnxxWYAp259BIygCUEdYjAQx8AXnxJfCEyXyV-AD4VSyt8b1xYNgAeQiYANwjgEJh5JIB6VIyLDXliq3jEji5eAWFRd08fH2L5EAAaEBs6EiYAcxQQJh5sPHxcAE9MPSkABW5epjoAeUxmWwVpGAgeLRCXAFpMecX97zJGfZthpkofAqY7c3V1SRi8vKvMG7I1ugBZCDEZD4VQgMiUSig9TyfBgH4PHoIMD4OZQBbLVYsOhgRRmdrgAAWEAA7gBJOi4Hx0cFgFDkMQIeRAA
### Repro steps
1. Write a render prop function which happens to contain nested functions with component-like names
2. Enable react compiler
3. App crashes due to hook errors because the render prop function got compiled. (This is expected behavior)
4. Try to fix the problem by applying 'use no memo' to the toplevel function
5. Nested function continues to be compiled; 'use no memo' needs to be applied to *all* nested component-like function names
As a side note, the [component-hook-factories](https://react.dev/reference/eslint-plugin-react-hooks/lints/component-hook-factories) rule does not detect this error: https://eslint-online-playground.netlify.app/#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
We ran into this while compiling a project that uses react-virtualized. react-virtualized takes a [render prop](https://github.com/bvaughn/react-virtualized/blob/c737715486f724586aee8870ebea1e9efb7b0bfe/docs/Table.md#headerrowrenderer) which gets called as a function rather than rendered as a component. Our render prop function crashed when it got compiled and passed into this prop.
### How often does this bug happen?
Every time
### What version of React are you using?
18.3.1
### What version of React Compiler are you using?
1.0.0 | open | yepitschunked | 2025-12-11T22:16:39 | 2025-12-11T22:16:39 | null | [
"Type: Bug",
"Status: Unconfirmed"
] | null | 0 | {
"heart": 2
} | 0 | false | 2026-03-29T04:51:59.996949 | |||
facebook | react | 16,173 | MDU6SXNzdWU0NzExNjE5NjE= | false | Why is react-dom so big? Why can't we tree-shake it? | **Do you want to request a *feature* or report a *bug*?**
Request a feature
**What is the current behavior?**
Bundling react-dom with the current CJS module is _HUGE_ - 124KB of stuff we don't use.
The bundle can be so much smaller by adding the source files into the npm package and enabling users to do a proper tree-shaking by defining the file in webpack
**If the current behavior is a bug, please provide the steps to reproduce and if possible a minimal demo of the problem. Your bug will get fixed much faster if we can run your code and it doesn't have dependencies other than React. Paste the link to your JSFiddle (https://jsfiddle.net/Luktwrdm/) or CodeSandbox (https://codesandbox.io/s/new) example below:**
**What is the expected behavior?**
When bundling with webpack, we should be able to properly tree-shake the unused code automatically.
**Which versions of React, and which browser / OS are affected by this issue? Did this work in previous versions of React?**
React 16.8.6
Browser Chrome 75
Linux | closed | completed | SGudbrandsson | 2019-07-22T15:29:33 | 2025-12-12T08:40:03 | 2019-07-22T17:42:58 | null | null | 0 | {
"+1": 20
} | 12 | false | 2026-03-29T04:52:00.061026 | ||
facebook | react | 35,355 | I_kwDOAJy2Ks7eFXXU | false | [Compiler Bug]:Compiler doesn't memoize if const is between hooks | ### What kind of issue is this?
- [x] React Compiler core (the JS output is incorrect, or your app works incorrectly after optimization)
- [ ] babel-plugin-react-compiler (build issue installing or using the Babel plugin)
- [ ] eslint-plugin-react-hooks (build issue installing or using the eslint plugin)
- [ ] react-compiler-healthcheck (build issue installing or using the healthcheck script)
### Link to repro
https://playground.react.dev/#N4lgzg9grgTgxgUxALhAejQAgLIILYQCWAXggCaYBGAnpgEolCGc
### Repro steps
1. Open the React Compiler Playground link: https://playground.react.dev/#N4lgzg9grgTgxgUxALhAejQAgLIILYQCWAXggCaYBGAnpgEolCGc
2. Observe the code that has a `const` declaration between React hooks
3. Check the compiled output - you'll see that memoization is not working correctly
4. The compiler should memoize values even when const is between hooks, but it currently doesn't
### How often does this bug happen?
Every time
### What version of React are you using?
Latest (React 19 or main branch)
### What version of React Compiler are you using?
Latest babel-plugin-react-compiler from main branch | closed | completed | pritjasani08 | 2025-12-13T13:10:35 | 2025-12-13T21:21:02 | 2025-12-13T21:20:43 | [
"Type: Bug",
"Status: Unconfirmed"
] | null | 0 | {} | 1 | true | spam | 2026-03-29T04:52:00.150355 | |
facebook | react | 35,352 | I_kwDOAJy2Ks7d63pg | false | [Compiler Bug]: Function hoisting ignored in eslint rule react-hooks/immutability | ### What kind of issue is this?
- [ ] React Compiler core (the JS output is incorrect, or your app works incorrectly after optimization)
- [ ] babel-plugin-react-compiler (build issue installing or using the Babel plugin)
- [x] eslint-plugin-react-hooks (build issue installing or using the eslint plugin)
- [ ] react-compiler-healthcheck (build issue installing or using the healthcheck script)
### Link to repro
https://playground.react.dev/#N4Igzg9grgTgxgUxALhAMygOzgFwJYSYAEAsgJ4DCEAtgA6EKY4AUtMEtYAlEcADrEiUMAgCiaNAlzNmPALwA+XgKKqiaCBFkBuFUQC+XXYID0J9VlwFiAEykAbAIYxH+QmCLOERBBKn4ANwR7Mk8cIhwAC28cDiIINAjoojA4DgQ9DGw3Yg0tHmADIj0YBBxYYkwoe3tjfQEQABoQNMw0PABzFBA8OggYcJwyWm9CgAV7KA68TAB5WhyPfXV2aiIAcgAjR03ggFpaSenMPdLHXD20ujx7BBgTGzwwHHXjAWZ+UxMr2hvXaxIEDsyGKIEcNT4IAEyzA-zA7QQHgmUxm80WRia4EiEAA7gBJJh3TDgsAoNAkhD6IA
### Repro steps
eslint rule `react-hooks/immutability` ignores [function hoisting](https://developer.mozilla.org/en-US/docs/Web/JavaScript/Guide/Functions#function_hoisting). This
```
function MyComponent(props) {
useEffect(() => {
foo();
});
// function declarations are effectively at the top of the scope
function foo() { }
return null;
}
```
should not cause error
```
Error: Cannot access variable before it is declared
`foo` is accessed before it is declared, which prevents the earlier access from updating when this value changes over time.
```
### How often does this bug happen?
Every time
### What version of React are you using?
19.2
### What version of React Compiler are you using?
1.0 | closed | completed | nifgraup | 2025-12-12T12:09:38 | 2025-12-15T09:17:33 | 2025-12-12T22:25:01 | [
"Type: Bug",
"Status: Unconfirmed"
] | null | 0 | {} | 2 | false | 2026-03-29T04:52:00.232116 | ||
facebook | react | 31,356 | I_kwDOAJy2Ks6b3Dcp | false | [React 19] Parameter order for server action when using `useActionState` | ## Summary
`useActionState` currently requires accepting an additional state argument as the _first_ parameter of a server action, like so (adapted from [Next.js documentation](https://nextjs.org/docs/app/building-your-application/data-fetching/server-actions-and-mutations#server-side-form-validation)):
```diff
// in app/actions.ts
-export async function updateUser(userId: string, formData: FormData) {
+export async function updateUser(prevState: any, userId: string, formData: FormData) {
```
This interferes with the use case of binding leading arguments to the action (adapted from [Next.js documentation](https://nextjs.org/docs/app/building-your-application/data-fetching/server-actions-and-mutations#passing-additional-arguments)):
```typescript
// in app/client-component.tsx
import { updateUser } from './actions'
const updateUserWithId = updateUser.bind(null, 'some id')
const [state, formAction] = useActionState(updateUserWithId, 'initial state')
```
For the above example to work, `updateUser` would need to accept the previous state as the _middle_ parameter:
```diff
// in app/actions.ts
-export async function updateUser(prevState: any, userId: string, formData: FormData) {
+export async function updateUser(userId: string, prevState: any, formData: FormData) {
```
Additionally, the server action might not actually care about the previous state. For example, if the action validates form data and the state is just a validation error message, then the action likely doesn't care about previous error messages. In such cases, it would be nicer DX to be able to omit the state parameter.
## Proposal
Make `useActionState` pass the previous state to the server action as the _last_ argument, always:
```diff
// in app/actions.ts
-export async function updateUser(prevState: any, userId: string, formData: FormData) {
+export async function updateUser(userId: string, formData: FormData, prevState: any) {
```
This might also enable skipping the serialization of the previous state when the server action does not use it (i.e. when `action.length` < 2).
| closed | not_planned | jonathanhefner | 2024-10-25T18:59:45 | 2025-12-14T14:04:59 | 2025-12-14T14:04:59 | [
"Resolution: Stale",
"React 19"
] | null | 0 | {
"-1": 1,
"eyes": 6
} | 13 | false | 2026-03-29T04:52:00.336189 | ||
facebook | react | 24,740 | I_kwDOAJy2Ks5L7b-U | false | Bug: `hidden` attribute does not accept string values | <!--
Please provide a clear and concise description of what the bug is. Include
screenshots if needed. Please test using the latest version of the relevant
React packages to make sure your issue has not already been fixed.
-->
React version:
```
"react": "18.3.0-next-522f47345-20220614",
"react-dom": "18.3.0-next-522f47345-20220614",
```
## Steps To Reproduce
```jsx
<span hidden="until-found">Hello React<span>
```
<!--
Your bug will get fixed much faster if we can run your code and it doesn't
have dependencies other than React. Issues without reproduction steps or
code examples may be immediately closed as not actionable.
-->
Link to code example:
https://codesandbox.io/s/headless-violet-ygn593
## The current behavior
React will turn hidden into a boolean attribute and removes "until-found":
## The expected behavior
React does not strip "until-found" to allow using [hidden-until-found in Chrome 102+](https://developer.chrome.com/articles/hidden-until-found/) for better SEO and accessibility. | open | jantimon | 2022-06-16T17:25:12 | 2025-12-12T21:31:27 | null | [
"Type: Bug",
"Component: DOM",
"Status: Unconfirmed"
] | null | 0 | {
"+1": 52
} | 6 | false | 2026-03-29T04:52:00.435188 | |||
facebook | react | 28,539 | I_kwDOAJy2Ks6B5aM_ | false | Bug: ESLint rule doesn't catch mistake with non-exhaustive dependencies | React version: 18.2
## Steps To Reproduce
1. Use the `useState()` hook.
* For example: `const [count, setCount] = useState(0)`
2. Create a function that directly uses the setter.
* For example: `const inc: () => setCount((count) => count + 1)`
3. Witness how the function using the setter is not required to be listed as dependency, leading to soundness issues.
* For example:
```jsx
const incAndRemember = useCallback(() => {
inc();
setLast({ last: inc }); // We might be setting a stale reference here,
}, []); // because `inc` doesn't need to be specified here.
```
4. Behavior in other hooks will start to diverge from what was intended if we ever rely on the identity of the reference we set.
Link to code example: https://github.com/arendjr/react-unstable-handler
## The current behavior
When a function inside a component (only) uses a setter of a `useState()` hook, that function is determined to be "stable" (non-reactive) despite its identity being unstable. Because the ESLint rule assumes the function is stable, it doesn't require it to be listed in dependency arrays, which may lead to soundness issues with other hooks, such as witnessed in the linked repository.
It appears the current behavior is in violation of the documentation regarding reactive values (https://react.dev/learn/lifecycle-of-reactive-effects#all-variables-declared-in-the-component-body-are-reactive ), which says:
> **All values inside the component (including props, state, and variables in your component’s body) are reactive. Any reactive value can change on a re-render, so you need to include reactive values as Effect’s dependencies.**
## The expected behavior
Whether a function is considered stable or not should be strictly determined by its identity.
| closed | not_planned | arendjr | 2024-03-11T14:40:57 | 2025-12-14T18:07:48 | 2025-12-14T18:07:48 | [
"Status: Unconfirmed",
"Resolution: Stale"
] | null | 0 | {} | 14 | false | 2026-03-29T04:52:00.528809 | ||
facebook | react | 32,290 | I_kwDOAJy2Ks6oY4XP | false | Bug: pre-trigger form submit in multistep form | There is multistep form with 3 steps: 1st, 2nd steps don't contain forms and 3rd contains. There is general submit button for all steps: it is located outside steps and connected with 3rd form by its id (button has "form" attribute with the same id as form). Value for button "form" attribute is changing depending on active step.
React version: 18.2.0
## Steps To Reproduce
1. Open sandbox console
2. Click submit button 2 times in order to transit to 3rd step
3. See that onSubmit handler has already fired
Link to code example: https://codesandbox.io/p/sandbox/pre-trigger-form-validation-react-18-cc3xl5
## The current behavior
Form of 3rd step has submitted immediately after transition from 2nd step
## The expected behavior
Form of 3rd step will not be submitted immediately after transition from 2nd step and will be submitted only after click submit btn which is connected with this form by id | closed | not_planned | evgeniyworkbel | 2025-02-01T09:15:29 | 2025-12-13T23:05:06 | 2025-12-13T23:05:06 | [
"Status: Unconfirmed",
"Resolution: Stale"
] | null | 0 | {} | 12 | false | 2026-03-29T04:52:00.579070 | ||
facebook | react | 34,416 | I_kwDOAJy2Ks7KLddr | false | Bug: webview tag not properly rendering allowpopups boolean attribute? | React version: v18.2.0
## Steps To Reproduce
Setup a \<webview\> element and add an allowpopups={true} and you would expect this allowpopups element to be added to the rendered element. It is not.
### Code:
<img width="388" height="163" alt="Image" src="https://github.com/user-attachments/assets/a1bc94d0-206e-4920-b0f2-814d15d962ec" />
### Rendered Elements from Devtools:
<img width="593" height="94" alt="Image" src="https://github.com/user-attachments/assets/b1d0bad3-fbb1-47d3-ba3a-228729d9a591" />
### The WebViewHTMLAttributes Interface
<img width="596" height="343" alt="Image" src="https://github.com/user-attachments/assets/9fef61fb-d1de-4e08-a72b-2edf46f643a5" />
## Example Repo
https://github.com/halo3consulting/react-webview-allowpopups-attribute-example
```
git clone https://github.com/halo3consulting/react-webview-allowpopups-attribute-example
cd react-webview-allowpopups-attribute-example
npm install
npm run dev
```
Expand the dev-tools elements to look at the rendered webview. You can see that the allowpopups element is not there.
## The current behavior
The webview element does not render with the allowpopups attribute when the value is true.
## The expected behavior
The webview element does render with the allow popups attribute.
| closed | not_planned | jamesrusso | 2025-09-07T21:03:26 | 2025-12-15T13:18:25 | 2025-12-15T13:18:25 | [
"Status: Unconfirmed",
"Resolution: Stale"
] | null | 0 | {
"heart": 1
} | 6 | false | 2026-03-29T04:52:00.654052 | ||
facebook | react | 35,367 | I_kwDOAJy2Ks7ehwRA | false | Improve cyclic thenable detection in ReactFlightReplyServer | Closing this PR for now; I want to rethink the approach and will open a new one if needed. | closed | completed | cvan20191 | 2025-12-16T06:28:41 | 2025-12-16T06:35:05 | 2025-12-16T06:32:09 | [
"Status: Unconfirmed"
] | null | 0 | {} | 0 | false | 2026-03-29T04:52:00.746871 | ||
facebook | react | 35,291 | I_kwDOAJy2Ks7cXJty | false | Bug: set-state-in-use-effect false negative when there are multiple instances | This code correctly produces a lint error:
```tsx
function MyComponent() {
const [loading, setLoading] = useState(false);
useEffect(() => {
setLoading(true); // ✅ ESLint correctly flags this (1 error)
}, []);
// commented out code...
return null;
}
```
But this code does not:
```tsx
function MyComponent() {
const [loading, setLoading] = useState(false);
useEffect(() => {
setLoading(true); // ❌ BUG: Error no longer reported (should still be flagged!)
}, []);
const [firstName, setFirstName] = useState<string>('');
const [lastName, setLastName] = useState<string>('');
const [fullName, setFullName] = useState<string>('');
useEffect(() => {
setFullName(`${firstName} ${lastName}`); // ❌ BUG: Also not reported
}, [firstName, lastName]);
return null;
}
```
React version: `^19.2.1`
## Steps To Reproduce
Created a minimal reproduction repo: https://github.com/roryabraham/set-state-in-effect-bug
## The current behavior
In the second code snippet provided, no lint errors are thrown.
## The expected behavior
Lint errors should be thrown for both `setState` calls in `useEffect` callbacks. | closed | completed | roryabraham | 2025-12-05T00:26:04 | 2025-12-16T08:21:35 | 2025-12-06T19:30:56 | [
"Status: Unconfirmed",
"Component: React Compiler",
"Resolution: Needs More Information"
] | null | 0 | {
"heart": 3
} | 4 | false | 2026-03-29T04:52:00.786677 | ||
facebook | react | 29,037 | I_kwDOAJy2Ks6IaNjd | false | [React 19] custom element property vs. attribute behavior deviates from RFC discussion and proposal doc | ## Summary
Hey folks! Not sure if the decisions here changed and I missed it, but it looks like the current React 19 beta impl for react props -> web component props/attrs has changed from what was originally [discussed](https://docs.google.com/document/d/1wwL_YZ5TnYorEEr2Lkeh9rgbnsynDXXI7tCejVcpE6I/edit#heading=h.86shuzhenihc) by @josepharhar. The tl;dr:
Unlike Preact, "complex values" were originally going to only be assigned to properties (if/when they exist). If the property does not exist, no value would be assigned. Instead, [only functions and Symbols are omitted from setting as attributes](https://github.com/facebook/react/blob/main/packages/react-dom-bindings/src/client/DOMPropertyOperations.js#L111-L114).
Not only does this result in likely undesired behavior (e.g. `anattr="[object Object]"`), it can also result in runtime errors that are potentially difficult to debug, e.g.:
```tsx
<my-wc iThoughtThisWasAProp={[aSymbol]}><my-wc> // This will crash
```
It looks like the originally planned logic could be introduced [in this general vicinity](https://github.com/facebook/react/blob/main/packages/react-dom-bindings/src/client/DOMPropertyOperations.js#L224-L227). Not sure if this is a bug, a feature request, or just a discussion as to why the change in plans, but just wanted to surface for visibility, just in case. Also, happy to cobble together a codesandbox or similar if that would be helpful!
<!--
Please provide a CodeSandbox (https://codesandbox.io/s/new), a link to a
repository on GitHub, or provide a minimal code example that reproduces the
problem. You may provide a screenshot of the application if you think it is
relevant to your bug report. Here are some tips for providing a minimal
example: https://stackoverflow.com/help/mcve.
-->
| closed | not_planned | cjpillsbury | 2024-05-09T22:41:16 | 2025-12-15T16:07:59 | 2025-12-15T16:07:59 | [
"Resolution: Stale",
"React 19"
] | null | 0 | {
"+1": 2
} | 23 | false | 2026-03-29T04:52:00.979960 | ||
facebook | react | 32,965 | I_kwDOAJy2Ks6zOEjz | false | Bug: Browser builds of `/static` and `/server` hang indefinitely in Node upon completion in v19 | When running in Node, `renderToString` (along with all other rendering methods, from my tests) from `react-dom/server.browser` and `prerender` from `react-dom/static.browser` seem to hang indefinitely -- the actual rendering completes just fine but the Node process is completely stuck afterwards and will not exit. This was not the case in v18, `server.browser` worked just fine in Node there.
React version: 19.1.0
## Steps To Reproduce
```js
// index.js
import { jsx } from 'react/jsx-runtime';
import { renderToString } from 'react-dom/server.browser';
const App = () => jsx("h1", { children: "Hello World!" });
const content = renderToString(jsx(App, {}));
console.log(content); // Prints `<h1>Hello World!</h1>` as expected, no issue here
```
1. `$ node ./index.js`
Code example: https://stackblitz.com/edit/node-6nxmucbr?file=index.js
## The current behavior
Hangs indefinitely, process will not exit once prerendering has completed
## The expected behavior
Process should exit once prerendering has completed | closed | not_planned | rschristian | 2025-04-20T05:09:50 | 2025-12-16T17:51:45 | 2025-12-16T16:57:26 | [
"Status: Unconfirmed"
] | null | 0 | {
"+1": 4,
"eyes": 2
} | 11 | false | 2026-03-29T04:52:01.082773 | ||
facebook | react | 14,830 | MDU6SXNzdWU0MDkwMDcwOTk= | false | Prop updates updating state in hooks | <!--
Note: if the issue is about documentation or the website, please file it at:
https://github.com/reactjs/reactjs.org/issues/new
-->
**Do you want to request a *feature* or report a *bug*?**
More of a question / discussion as I could not find it in the documentation.
**What is the current behavior?**
Currently when using hooks like `useState` and `useReducer`, any updates to props used by their initial state has no effect on the state managed by these hooks. For instance, consider this example from the Hooks API Reference:
```
function Counter({initialCount}) {
const [count, setCount] = useState(initialCount);
return (
<>
Count: {count}
<button onClick={() => setCount(initialCount)}>Reset</button>
<button onClick={() => setCount(prevCount => prevCount + 1)}>+</button>
<button onClick={() => setCount(prevCount => prevCount - 1)}>-</button>
</>
);
}
```
Now let's suppose that a parent component initially renders a `Counter` with `initialCount` set to `0`. But later on the `initialCount` prop passed by the parent component is updated to let's say `11`. I would expect that a change in the `initialCount` prop, would then trigger a change in the `count` state but that doesn't happen.
At first I wasn't sure why the `count` was not updating as I expect any data changes to flow through where they are used. It seems like that the initial state is only used *initially* as the name suggests in `useState` or `useReducer` and any changes to it are ignored by these hooks.
However, we are still able to address the problem by using effects:
```
function Counter({initialCount}) {
const [count, setCount] = useState(initialCount);
useEffect(() => setCount(initialCount), [initialCount]);
return (
<>
Count: {count}
<button onClick={() => setCount(initialCount)}>Reset</button>
<button onClick={() => setCount(prevCount => prevCount + 1)}>+</button>
<button onClick={() => setCount(prevCount => prevCount - 1)}>-</button>
</>
);
}
```
The need to use effects wasn't immediate to me but it makes sense as we would generally *hook* such behavior into component lifecycle methods like `componentDidUpdate`. I am curious as to why this doesn't already happen automatically through `useState` and `useReducer`, and wondering if this is the correct way to listen to such prop changes and accordingly update state, or is there some better way?
We could also extract this logic into a custom hook:
```
function useStateWithEffect(initialState) {
const [state, setState] = useState(initialState);
useEffect(() => setState(initialState), [initialState]);
return [state, setState]
}
// Haven't tested the reducer but as an idea
function useReducerWithEffect(reducer, initialArg, init = x => x) {
const reducerWithStateUpdate = (state, action) => action.type === 'updateInitialArg' ? init(action.initialArg) : reducer(state, action);
const [state, dispatch] = useReducer(reducerWithStateUpdate, initialArg, init);
useEffect(() => dispatch({ type: 'updateInitialArg', initialArg }), [initialArg]);
return [state, dispatch]
}
```
I spent some time trying to figure this out so if someone has reference to material already discussing this problem, please do share as I couldn't find it while going through the docs / posts about hooks. | closed | completed | iRule95 | 2019-02-11T22:30:38 | 2025-12-16T22:01:45 | 2019-02-12T01:13:37 | null | null | 0 | {
"+1": 27,
"rocket": 2
} | 12 | false | 2026-03-29T04:52:01.351214 | ||
facebook | react | 33,015 | I_kwDOAJy2Ks6z5v5K | false | Bug: ViewTransition name prop does not support numerical prefixes | `ViewTransition` does not animate if a `name` prop starting with a number is provided. Presumably this is because the name is being used as a CSS class, but that is not obvious from the documentation.
React version: experimental
## Steps To Reproduce
1. Create a ViewTransition with the name "1name" that wraps an element
2. Insert or remove the ViewTransition
Link to code example: https://codesandbox.io/p/sandbox/hopeful-dawn-cfgv64
## The current behavior
Animation does not fire and no errors or warnings present
## The expected behavior
One of the following:
1. The ViewTransition applies (e.g. by escaping the name)
2. An warning is logged for invalid ViewTransition name
3. The React docs are updated to call out that name must be a valid CSS name
| closed | not_planned | roarosa | 2025-04-24T19:29:41 | 2025-11-19T01:33:50 | 2025-11-19T01:33:50 | [
"Status: Unconfirmed",
"Resolution: Stale"
] | null | 0 | {
"+1": 3
} | 6 | false | 2026-03-29T04:51:55.461173 | ||
facebook | react | 35,138 | I_kwDOAJy2Ks7YLcAk | false | [Compiler Bug]: stable setter for useOptimistic | ### What kind of issue is this?
- [x] React Compiler core (the JS output is incorrect, or your app works incorrectly after optimization)
- [ ] babel-plugin-react-compiler (build issue installing or using the Babel plugin)
- [ ] eslint-plugin-react-hooks (build issue installing or using the eslint plugin)
- [ ] react-compiler-healthcheck (build issue installing or using the healthcheck script)
### Link to repro
https://playground.react.dev/#N4Igzg9grgTgxgUxALhAMygOzgFwJYSYAEAwhALYAOhCmOAFMGAgDYK4CCuBmAvgJRFgAHUzCccQmBxEA2njAAFWgBM8mAOYAaIsxwBJJavUaAukQC8RKMwDylfOQX449NAEMWzfgG5R-iSkZd25CS2tmEk8WACMQgGt6ekELAD4hAPE9Q2VMNU16HBgoBF9MooQcWGJmNk5QzGS-MRxeHVla9hwufEJTMrEWmErqonpygB5U8vEJgGVWLv1MSihghotgEN6+IgB6aZbxcWAFXPyNIgAyK6IJsEp3TFSAcgA6D5eJvYen1N5JgcAvxRLwQFoQJJMGg8BoUCA8FQIDAZDgAJ6UBBCIiKFhQDTqew7MBEXhENAwChEF5xGKsAC0lDxBMw9OG23pkioeDYMD2amkL2aokYoiI+z2XMoPPcOwAshAVAhkERhCBomqxaqQI9MHg4AAVAAWwzARogLBUapVaswNDVWi1apw7hgGkq1u1AEYAJwO0G6WUKGEIEm4-GEhw8MC+cHgc0Ad2WOAQMEwnjAKA8XgQvCAA
### Repro steps
<img width="1280" height="635" alt="Image" src="https://github.com/user-attachments/assets/47665500-8654-4d6b-9d20-444460671c2f" />
### How often does this bug happen?
Every time
### What version of React are you using?
19.2
### What version of React Compiler are you using?
playground | closed | completed | rickhanlonii | 2025-11-14T19:24:08 | 2025-11-18T15:50:44 | 2025-11-18T15:50:44 | [
"Type: Bug"
] | null | 0 | {} | 1 | false | 2026-03-29T04:51:55.535670 | ||
facebook | react | 35,163 | I_kwDOAJy2Ks7YyNjk | false | [Compiler Bug]: React Compiler infers incorrect dependencies based on variable declaration order | ### What kind of issue is this?
- [x] React Compiler core (the JS output is incorrect, or your app works incorrectly after optimization)
- [ ] babel-plugin-react-compiler (build issue installing or using the Babel plugin)
- [ ] eslint-plugin-react-hooks (build issue installing or using the eslint plugin)
- [ ] react-compiler-healthcheck (build issue installing or using the healthcheck script)
### Link to repro
https://playground.react.dev/#N4Igzg9grgTgxgUxALhHCA7MAXABAWwEMAHXAXlwwQHdcBZEgCgG0AdDXXNkQ1kAGlzACASwzJcARgC+AXXaduAIz6Dh+MRIBMc9rICU7dggAexCDDzoseAHJR8CGCLgBJDMSjYAwhHzmqDDwKRmEAN0IAGygEQUwANSiY7wALQgwAcwQAE1xpfXIAPiEFXGscSgclJwAVAE9iBHJcCOiEAH4AOizse3xqmHrGxn0AbiMOTjLMCoSkhFT0rOzvKMilQjgAa2bGKmpEtokMKqcCsmLgUqmpubbFzJzGQ5jOgDMYPz6BvZoXhH0Y2ueXGGGB5Tw6jEeWaRGI3QQ2D2p0GDQBAEJQcCYIjYBxGMCbgAePpOFzuTzYQk3KatGJkYB0joI3ookbSak00QYBkaDAcyZcm53ZJpR7ZBkihZi5arSLrTZbAVC3AAekKwKB-NBAjQmDeIgyKBAIn8Fjw2DRQlwAAVohkxAB5YjYEQzGEfPy4ADkG2qkQAtMR7WIAzjNtgA+h-CJIk5VdkRDhvViMKFSqrVdHiLHCK7MHQINkEBI+Gs+OxpLgwHmkwaEGBbSGMM781gxrqwCkINR3NgnBgomAUG8hwhpEA
### Repro steps
The React Compiler infers incorrect dependencies for the variable min when its declaration appears after the callback definition. Given the following code:
``` typescript
const map = new Map([
["a", { min: 1 }],
["b", { min: 2 }],
]);
export const NumericInputComponent = ({ value, onValueChanged }) => {
const numberType = value?.getNumberType();
const onValueChangedCallback = (newValue: number) => {
onValueChanged(Value.fromNumber(newValue));
};
const { min } = map.get(numberType)!;
return (
<NumericInput
value={value?.getNumber()}
min={min}
onValueChanged={onValueChangedCallback}
/>
);
};
```
In the compiled output, the compiler reports that min depends on both value and onValueChanged.
This is incorrect, because min is derived exclusively from numberType → value, and does not use onValueChanged in any way.
If the declarations are reversed:
``` typescript
const { min } = map.get(numberType)!;
const onValueChangedCallback = (newValue: number) => {
onValueChanged(Value.fromNumber(newValue));
};
```
…then the compiler correctly infers that min depends only on value.
Expected behavior:
min should depend solely on value, regardless of declaration order.
Actual behavior:
min incorrectly depends on onValueChanged when declared after the callback.
Impact:
This causes unnecessary reactivity and avoidable re-renders.
### How often does this bug happen?
Every time
### What version of React are you using?
19.2
### What version of React Compiler are you using?
1 | closed | completed | nicolo-danzi-ok | 2025-11-18T09:24:13 | 2025-11-19T05:12:07 | 2025-11-19T05:12:07 | [
"Type: Bug",
"Status: Unconfirmed"
] | null | 0 | {} | 1 | false | 2026-03-29T04:51:55.573216 | ||
facebook | react | 30,691 | I_kwDOAJy2Ks6S9uHH | false | [React 19] No re-render after 'useActionState' action queue finishes | ## Summary
When multiple client side actions are scheduled via `useActionState`, the "Action queue" promises are processed sequentially as expected. However after the last action promise resolves, the component is not re-rendered. This means, the component is stuck in "loading" without access to "data".
### Steps to reproduce
1. Open the demo here: https://codesandbox.io/p/sandbox/use-action-state-stuck-xl72xk?file=%2Fsrc%2FApp.js
2. Click the "Send request" button two times.
3. After 10 seconds (each request is 5 seconds and processed sequentially), the component still shows "Loading..." and not the dummy data (as I would expect).
#### Notes
* When only one "action" is scheduled (button clicked once), the component re-renders when the action is done, as expected.
* The promise "delay" seems to have an effect. When `REQUEST_DELAY` is set lower i.e. 1000ms, this "issue" is not present.
* Possibly related: https://github.com/facebook/react/issues/27630
Is this behavior intentional?
| open | morditore | 2024-08-14T12:04:29 | 2025-11-18T16:49:26 | null | [
"Type: Bug",
"React 19"
] | null | 0 | {
"+1": 4
} | 7 | false | 2026-03-29T04:51:55.638267 | |||
facebook | react | 32,196 | I_kwDOAJy2Ks6nSgtC | false | [DevTools Bug]: [DevTools Bug] Cannot reorder children for node "0" because no matching node was found in the Store. | ### Website or app
https://www.autotrack.nl/aanbod
### Repro steps
1. Start profiling in chrome
2. Click on any filter
3. Uncaught Error: Cannot reorder children for node "0" because no matching node was found in the Store.
Dismiss
The error was thrown at chrome-extension://fmkadmapgofadopljbjfkapdkoienihi/build/main.js:1:1175534
at v.emit (chrome-extension://fmkadmapgofadopljbjfkapdkoienihi/build/main.js:1:1140783)
at chrome-extension://fmkadmapgofadopljbjfkapdkoienihi/build/main.js:1:1142390
at bridgeListener (chrome-extension://fmkadmapgofadopljbjfkapdkoienihi/build/main.js:1:1552529)
### How often does this bug happen?
Every time
### DevTools package (automated)
react-devtools-extensions
### DevTools version (automated)
6.0.1-c7c68ef842
### Error message (automated)
Cannot reorder children for node "0" because no matching node was found in the Store.
### Error call stack (automated)
```text
at chrome-extension://fmkadmapgofadopljbjfkapdkoienihi/build/main.js:1:1175534
at v.emit (chrome-extension://fmkadmapgofadopljbjfkapdkoienihi/build/main.js:1:1140783)
at chrome-extension://fmkadmapgofadopljbjfkapdkoienihi/build/main.js:1:1142390
at bridgeListener (chrome-extension://fmkadmapgofadopljbjfkapdkoienihi/build/main.js:1:1552529)
```
### Error component stack (automated)
```text
```
### GitHub query string (automated)
```text
https://api.github.com/search/issues?q=Cannot reorder children for node because no matching node was found in the Store. in:title is:issue is:open is:public label:"Component: Developer Tools" repo:facebook/react
``` | open | MikeEzzati | 2025-01-23T11:25:04 | 2025-11-19T03:33:40 | null | [
"Type: Bug",
"Status: Unconfirmed",
"Component: Developer Tools"
] | [
"hoxyq"
] | 0 | {
"+1": 10
} | 5 | false | 2026-03-29T04:51:55.744435 | |||
facebook | react | 35,171 | I_kwDOAJy2Ks7ZC9TS | false | [Compiler Bug]: `using` syntax is not preserved | ### What kind of issue is this?
- [x] React Compiler core (the JS output is incorrect, or your app works incorrectly after optimization)
- [ ] babel-plugin-react-compiler (build issue installing or using the Babel plugin)
- [ ] eslint-plugin-react-hooks (build issue installing or using the eslint plugin)
- [ ] react-compiler-healthcheck (build issue installing or using the healthcheck script)
### Link to repro
https://github.com/cormacrelf/react-compiler-using-bug
### Repro steps
`pnpm i && pnpm dev`. Open App.tsx to see there are three [diposables with `using` syntax](https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Statements/using). Open the console, note that only one of the Disposables had its `Symbol.dispose` method called.
To show that it's react-compiler's fault, you can:
- observe that only `using` syntax inside the component is transformed, the one at the top level is preserved.
- uncomment `effect.useIt()` which makes react-compiler skip the component and all the disposals work.
- skip the component
- disable the transform in vite.config.ts
Note that react-compiler transforms the `using render = new Disposable("render");` to just `new Disposable("render")`, which is not a valid transform. It would be valid for a `const` or `let` binding.
There is another bug in here that I can't reproduce, because react-compiler keeps bailing out when I add the `useIt` call. But very similar code to the following *does* get compiled, and in that case the using is just transformed to a const, which again breaks the disposal.
```typescript
function App() {
const context = useContext();
const ref = React.useRef(null);
useEffect(() => {
if (ref.current) {
using effect = new Disposable(context);
effect.useIt();
}
});
return <div ref={ref}></div>;
}
```
### How often does this bug happen?
Every time
### What version of React are you using?
19.2.0
### What version of React Compiler are you using?
1.0.0 | open | cormacrelf | 2025-11-19T07:29:04 | 2025-11-19T22:26:25 | null | [
"Type: Bug",
"Status: Unconfirmed"
] | null | 0 | {} | 0 | false | 2026-03-29T04:51:55.781837 | |||
facebook | react | 35,209 | I_kwDOAJy2Ks7aX0NF | false | Bug: | <!--
Please provide a clear and concise description of what the bug is. Include
screenshots if needed. Please test using the latest version of the relevant
React packages to make sure your issue has not already been fixed.
-->
React version:
## Steps To Reproduce
1.
2.
<!--
Your bug will get fixed much faster if we can run your code and it doesn't
have dependencies other than React. Issues without reproduction steps or
code examples may be immediately closed as not actionable.
-->
Link to code example:
<!--
Please provide a CodeSandbox (https://codesandbox.io/s/new), a link to a
repository on GitHub, or provide a minimal code example that reproduces the
problem. You may provide a screenshot of the application if you think it is
relevant to your bug report. Here are some tips for providing a minimal
example: https://stackoverflow.com/help/mcve.
-->
## The current behavior
## The expected behavior
| closed | completed | RafaelHDSV | 2025-11-25T16:04:41 | 2025-11-25T16:04:50 | 2025-11-25T16:04:50 | [
"Status: Unconfirmed"
] | null | 0 | {} | 0 | false | 2026-03-29T04:51:55.919386 | ||
facebook | react | 35,183 | I_kwDOAJy2Ks7ZgatI | false | Bug: | <!--
Please provide a clear and concise description of what the bug is. Include
screenshots if needed. Please test using the latest version of the relevant
React packages to make sure your issue has not already been fixed.
-->
React version:
## Steps To Reproduce
1.
2.
<!--
Your bug will get fixed much faster if we can run your code and it doesn't
have dependencies other than React. Issues without reproduction steps or
code examples may be immediately closed as not actionable.
-->
Link to code example:
<!--
Please provide a CodeSandbox (https://codesandbox.io/s/new), a link to a
repository on GitHub, or provide a minimal code example that reproduces the
problem. You may provide a screenshot of the application if you think it is
relevant to your bug report. Here are some tips for providing a minimal
example: https://stackoverflow.com/help/mcve.
-->
## The current behavior
## The expected behavior
| closed | completed | luisfernando30077-lgtm | 2025-11-20T21:26:40 | 2025-11-24T16:37:34 | 2025-11-24T16:37:34 | [
"Status: Unconfirmed"
] | null | 0 | {} | 2 | false | 2026-03-29T04:51:55.947513 | ||
facebook | react | 16,773 | MDU6SXNzdWU0OTMxMTYyOTY= | false | [Concurrent Mode] App stops updating/rendering on inactive tab | **Do you want to request a *feature* or report a *bug*?**
Bug
**What is the current behavior?**
1. When Concurrent Mode is enabled
2. and the Chrome tab is inactive
3. and a new render happens inside app
4. the app won't render this update until I open the chrome tab.
This only happens on Concurrent Mode. I noticed this because I update the document title on a react Context, like this:
<img width="140" alt="Screen Shot 2019-09-12 at 23 58 05" src="https://user-images.githubusercontent.com/619186/64834916-5e0c2b80-d5ba-11e9-8d1c-d3581c9c237e.png">
and on Concurrent Mode it was never updating when the tab was inactive. Only when I opened the tab.
Disabling Concurrent Mode was enough for it to start working as expected.
I believe this is related to the Chrome throttling behavior of inactive tabs. For example, Chrome doesn't resolve neither `requestIdleCallback` neither `requestAnimationFrame` when the tab is inactive. Made this non-react example just to show this Chrome behavior because I believe it's related: https://codesandbox.io/embed/dank-fire-0zjxv
Reproduction using `useEffect` and allow changing between legacy/sync/concurrent modes: https://codesandbox.io/embed/react-bug-inactive-tabs-tnosy
### Incorrect behavior (sync or concurrent mode, it get's completely suspended)
**What is the expected behavior?**
Inactive tabs should have the same behavior with or without Concurrent Mode. They should keep updating/rendering things (ok if a little bit throttled), but NOT completely pause doing things until the tab becomes active.
**Which versions of React, and which browser / OS are affected by this issue? Did this work in previous versions of React?**
React 16.9
Chrome 76
macOS | closed | completed | brunolemos | 2019-09-13T03:10:12 | 2025-11-21T05:09:42 | 2019-09-14T02:11:58 | [
"Type: Bug"
] | null | 0 | {
"+1": 4
} | 15 | false | 2026-03-29T04:51:56.030591 | ||
facebook | react | 29,111 | I_kwDOAJy2Ks6JKDE5 | false | [React 19] Support React Compiler with older React versions for libraries | @poteto asked me to create this issue for discussion.
It would be useful for library authors if the React Compiler could work with older versions of React. Libraries often need to support many different versions of React to avoid breaking older applications that haven't had a chance to update yet. That's the case for us with React Aria, which supports 16.8+. We can't really release a major version that bumps the requirement to React 19 because we have to support many older applications still using 16-18 that haven't prioritized updating yet. We can't realistically leave those applications behind, or maintain multiple major versions of React Aria ourselves, which means we can't use the compiler.
It would be amazing if the compiler could optionally generate regular `useMemo`, or there was a shim available for older versions of React (similar to the `useSyncExternalStore` shim that was previously released). This could be available specifically for library authors who need to support multiple versions of React, rather than applications (which should just update to React 19). | closed | completed | devongovett | 2024-05-16T18:51:12 | 2025-11-24T23:04:32 | 2024-11-18T15:07:10 | [
"Component: React Compiler",
"Resolution: Stale"
] | null | 0 | {
"+1": 2
} | 6 | false | 2026-03-29T04:51:56.076393 | ||
facebook | react | 35,206 | I_kwDOAJy2Ks7aTQ3Y | false | [Compiler Bug]: no-deriving-state-in-effects does not error if you add the state setter to the dependencies array | ### What kind of issue is this?
- [ ] React Compiler core (the JS output is incorrect, or your app works incorrectly after optimization)
- [ ] babel-plugin-react-compiler (build issue installing or using the Babel plugin)
- [x] eslint-plugin-react-hooks (build issue installing or using the eslint plugin)
- [ ] react-compiler-healthcheck (build issue installing or using the healthcheck script)
### Link to repro
https://playground.react.dev/#N4Igzg9grgTgxgUxALhAgHgBwjALgAgBMEAzAQygBsCSoA7OXASwjvwFkBPAQU0wApgRCHCgBbBHVxh8AX2T5ghEeMnSFZOp1kBKRQB02+OKzAEA2stESpYADJMzAGnxgEuACIqb0h2YC6+AC8hvhh+FBuAMq4ZLgI+AA8+ABKCGTKdJQ8MDBknEn4XtZqHu5kTJQyAD74dFSU+AB8+LX1lI0t-FaqtjoA3Iah+AD0I-gA5kwAbggyZPhzlExSi7k4w2lkjAB0kQgAoiQkCIz8-HpBLcDDYW6e3mr2jrjdj32DRrIulu-S-gMhkYxsI5nUIAQprN8AslisCAh1jBNuldvsjiczhdgtdbq53MVer4Xm8Sh9ht98L8ydIXPdCT5ngFAXRZIYQE4QCY6CQmBMUCAmGJsHh8LhOJgEkIAAqUKBTOgAeUwzFMcnwJBgEDE+AA5AAjMj6hCUAC0mDlCtNMFRuFNJmFlURI0IL11n0MgmGYwdmEqcRYdHYEGICn0IDIHXDw0k0yYWroPgUNyMYXD00jTEIcQQADkIFF3DEcwBJOgY07ScMKXAwKAICmGWSuANgXlg2XylbK1V0MADDngAAWEAA7mX4jA6JGwChyFUELIgA
### Repro steps
1. See playground - one useEffect gives a eslint error the other does not
### How often does this bug happen?
Every time
### What version of React are you using?
19.1.0
### What version of React Compiler are you using?
1.0.0 | closed | completed | lukpsaxo | 2025-11-25T10:38:58 | 2025-11-25T16:36:59 | 2025-11-25T16:36:59 | [
"Type: Bug",
"Status: Unconfirmed"
] | null | 0 | {} | 1 | false | 2026-03-29T04:51:56.122922 | ||
facebook | react | 11,566 | MDU6SXNzdWUyNzQzMDkzNjI= | false | [Umbrella] New algorithm for resuming interrupted work | *Resuming* is the ability to re-use fibers after they are interrupted by a higher-priority update. Take the following scenario: A component is updated at a normal, async priority. Before the update is finished processing, a higher-priority update is scheduled (let's say it's synchronous, though it could also be a higher-priority async update). The sync update *interrupts* the async update, leaving it unfinished. After the sync update finishes, we go back to processing the interrupted, async update. It's possible, and even likely, that the interrupted work wasn't touched by the sync work and can be *resumed* without starting over completely.
This is an important optimization for several async features we have in mind, including error handling, blockers, pre-rendering, and hidden priority.
We used to have an implementation of resuming that mostly worked but had some bugs. A few months ago, I spent some time [identifying the bugs using fuzz testing](https://github.com/facebook/react/pull/9952) and fixing them by iterating on the existing algorithm. I eventually got a [version working that passed all the tests](https://github.com/facebook/react/pull/9695). But even this version didn't have all of the features we wanted, and the algorithm seemed inherently flawed. So we decided it would be best to scrap the existing algorithm and revisit resuming in the future.
We now believe we have a better idea of how resuming should work. I'm going to split the work into multiple PRs, and use this issue to keep track of our progress.
My apologies if some of my descriptions are hard to follow. It can be difficult to describe without resorting to jargon. I'll iterate on this issue as I work.
Always reconcile against current child set (#11564)
---------------------------------------------------
This is a small refactor that reflects what we already do without resuming: the set we reconcile against is always the current set. In the reverted resuming algorithm, the set we reconcile against was sometimes a work-in-progress set, and there are a few code paths that are left over from that implementation.
Stash interrupted children
--------------------------
When cloning a work-in-progress fiber from current, and there is already an existing work-in-progress that was interrupted, stash the interrupted work-in-progress children (and corresponding fields) in case we can reuse them later. In begin phase, add an additional check to see if incoming props/state match the interrupted props/state. If so, bail out and re-use the interrupted children. If not, the interrupted children are no longer useful, because we're about to re-render the parent and overwrite them. (Unmounted fibers actually can be re-used even if we re-render the parent; see next step.)
This gets us back to the same functionality we had in the old resuming algorithm. We can now resume interrupted children if we come back to it at the same priority at which it was originally rendered. The main limitation is that the work is lost if the parent is re-rendered at a higher priority.
**Need a way to distinguish between a work-in-progress fiber and the "previous current" fiber*
Pool unmounted, interrupted children so they can resume even if parent re-renders at higher priority
------------------------------------------------------------------------------------
When a fiber is about to be re-rendered, and there are interrupted children that could not be reused, search through the interrupted children and find the ones that are unmounted (don't have an alternate). Stash the unmounted children in a separate set; they can be kept around indefinitely without being overwritten. This set acts like a pool of children. The next time the parent is re-rendered at the priority of the interrupted children, check the pool for matches before creating new fibers.
| open | acdlite | 2017-11-15T21:24:06 | 2025-11-21T07:13:34 | null | [
"Type: Umbrella",
"Component: Reconciler",
"React Core Team"
] | null | 0 | {
"+1": 4
} | 2 | false | 2026-03-29T04:51:56.173104 | |||
facebook | react | 35,219 | I_kwDOAJy2Ks7ajgex | false | Bug: [React 19 | Streaming SSR] Script tags injected via ChunkExtractor not appearing in final HTML when using renderToPipeableStream | Description
I am using React 19, Express, and Streaming SSR (renderToPipeableStream) along with @loadable/server ChunkExtractor to extract and inject script/style tags for client-side hydration.
The loadable-stats.json is generated correctly, and ChunkExtractor.getScriptTags() returns the expected <script> tags.
However, when I inject these <script> tags inside the onAllReady() callback of renderToPipeableStream, they never appear in the final browser HTML output. This results in no client bundle being loaded and therefore hydration never occurs.
This issue happens even though:
Scripts are extracted correctly
onAllReady() is called
res.write(scripts) is executed
SSR output is streamed successfully
React version: 19.2.0
## Steps To Reproduce
https://github.com/sandeep3690Butte/node-r19-ssr
1.git clone [<repo>](https://github.com/sandeep3690Butte/node-r19-ssr)
2.yarn install
3.yarn build
4.yarn start
5.Open the SSR page → scripts do not appear in view-source or DevTools Network panel.
Repository Structure
client/ → all client React code
server/ → Express SSR server
webpack.* → server/client configs for dev + prod
## The current behavior
The onAllReady() callback is triggered correctly
res.write() executes with no error
But no script tags appear in the final HTML
Client JS never loads, causing hydration to never run
No errors in server logs or browser console
## The expected behavior
The <script> tags returned by ChunkExtractor.getScriptTags() should be written to the response in onAllReady() so they load the client bundles required for hydration.
| open | sandeep3690Butte | 2025-11-26T09:49:50 | 2025-11-26T10:19:26 | null | [
"Type: Feature Request",
"Component: Server Rendering"
] | null | 0 | {
"+1": 1,
"eyes": 1,
"heart": 1
} | 1 | false | 2026-03-29T04:51:56.238717 | |||
facebook | react | 35,195 | I_kwDOAJy2Ks7Z7McU | false | Bug: useEffect cleanup function throws TypeError when returning null in React 19 | <!--
Please provide a clear and concise description of what the bug is. Include
screenshots if needed. Please test using the latest version of the relevant
React packages to make sure your issue has not already been fixed.
-->
React version:
## Steps To Reproduce
1.
2.
<!--
Your bug will get fixed much faster if we can run your code and it doesn't
have dependencies other than React. Issues without reproduction steps or
code examples may be immediately closed as not actionable.
-->
Link to code example:
<!--
Please provide a CodeSandbox (https://codesandbox.io/s/new), a link to a
repository on GitHub, or provide a minimal code example that reproduces the
problem. You may provide a screenshot of the application if you think it is
relevant to your bug report. Here are some tips for providing a minimal
example: https://stackoverflow.com/help/mcve.
-->
## The current behavior
## The expected behavior
| closed | duplicate | PraveenBoss45 | 2025-11-23T15:56:54 | 2025-11-26T10:57:54 | 2025-11-26T10:57:54 | [
"Status: Unconfirmed"
] | null | 0 | {} | 2 | false | 2026-03-29T04:51:56.279705 | ||
facebook | react | 30,368 | I_kwDOAJy2Ks6P8qYs | false | Bug: useFormStatus pending state is reset when component state is updated | If you have a component that relies on the `pending` return value of `useFormStatus`, the `pending` state will incorrectly reset to `false` if the component is updated due to a `useState` update. This does not happen if the `useState` hook is placed in a child component.
React version: 19.0.0-rc-512b09b2-20240718
## Steps To Reproduce
[codesandbox.io/p/sandbox/react-useformstatus-pending-reset-on-unrelated-state-update-m59zw8](https://codesandbox.io/p/sandbox/react-useformstatus-pending-reset-on-unrelated-state-update-m59zw8)
1. Create a react app that uses form actions (this uses NextJS starter code)
2. Have the form action delay for a set period of time before resolving
3. Create a child component that uses `useFormStatus().pending`. Have the component also use a `useState` hook that updates on an interval using `useEffect`.
4. Place this child component as a child of the `<form />` element
5. Verify that `useFormStatus().pending` is reset to false as soon as a call to `setState` happns
Example component
```
export default function SubmissionState() {
const formStatus = useFormStatus();
const [counter, setCounter] = useState(0);
useEffect(() => {
const id = setTimeout(() => {
if (!formStatus.pending) {
setCounter(0);
return;
}
setCounter(counter + 1);
}, 1000);
return () => clearTimeout(id);
}, [counter, formStatus.pending]);
return (
<div>
{formStatus.pending ? `Pending for ${counter} seconds` : "Not pending"}
</div>
);
}
```
this example works as expected, `useFormStatus().pending` state matches the server actions state
```
function SubmissionStateBody({ pending }: { pending: boolean }) {
const [counter, setCounter] = useState(0);
useEffect(() => {
if (!pending) {
setCounter(0);
return;
}
const id = setTimeout(() => {
setCounter(counter + 1);
}, 1000);
return () => clearTimeout(id);
}, [counter, pending]);
return (
<div>{pending ? `Pending for ${counter} seconds` : "Not pending"}</div>
);
}
export default function SubmissionStateCorrect() {
const formStatus = useFormStatus();
return <SubmissionStateBody pending={formStatus.pending} />;
}
```
Link to code example:
[Github repro using nextjs starter template](https://github.com/jatwood/repro-use-form-issue)
## The current behavior
The UI shows the submission as pending for the length of the server action.
## The expected behavior
The UI shows the submission as not-pending as soon as the counter is updated.
| closed | completed | jatwood | 2024-07-18T01:59:20 | 2025-11-26T05:30:31 | 2025-11-19T17:22:08 | [
"Type: Bug",
"React 19"
] | null | 0 | {
"+1": 12,
"eyes": 2
} | 14 | false | 2026-03-29T04:51:56.487987 | ||
facebook | react | 35,244 | I_kwDOAJy2Ks7bD2jf | false | [DevTools Bug]: Internal state change in a cousin component causing a Profiler re-render in another cousin component wrapped inside parent divs. | ### Website or app
https://github.com/mukesharyal/react-devtools-bug-report
### Repro steps
Steps for reproduction:
1. Clone the GitHub repo.
2. Run the dev server and open React Devtools.
3. You will see two components `ComponentA` and `ComponentB`.
4. Cause a re-render of `ComponentA` by clicking the button.
5. `ComponentB` will also be highlighted as being re-rendered and can also be seen in a Profiler session.
`ComponentA` and `ComponentB` are wrapped inside of two `div`s in the `App` component like this:
function App() {
return(
<div>
<div>
<ComponentA />
</div>
<div>
<ComponentB />
</div>
</div>
);
}
export default App;
`
### How often does this bug happen?
Every time
### DevTools package (automated)
_No response_
### DevTools version (automated)
_No response_
### Error message (automated)
_No response_
### Error call stack (automated)
```text
```
### Error component stack (automated)
```text
```
### GitHub query string (automated)
```text
``` | closed | completed | mukesharyal | 2025-11-28T15:09:57 | 2025-11-28T15:46:40 | 2025-11-28T15:46:40 | [
"Type: Bug",
"Status: Unconfirmed",
"Component: Developer Tools"
] | null | 0 | {} | 1 | false | 2026-03-29T04:51:56.559966 | ||
facebook | react | 31,388 | I_kwDOAJy2Ks6cbnQu | false | [compiler] Support annotating hook factories that produce stable hooks | ### What kind of issue is this?
- [ ] React Compiler core (the JS output is incorrect, or your app works incorrectly after optimization)
- [ ] babel-plugin-react-compiler (build issue installing or using the Babel plugin)
- [ ] eslint-plugin-react-compiler (build issue installing or using the eslint plugin)
- [ ] react-compiler-healthcheck (build issue installing or using the healthcheck script)
### Link to repro
nope
### Repro steps
The application I'm developing follows a specific pattern for accessing Redux store data based on parameters. Here's an example:
```jsx
const ProductTile = ({ productId }) => {
const useProductSelector = createUseSelectorWithParam(productId)
const title = useProductSelector(productStore.selectLabel)
const value = useProductSelector(productStore.selectValue)
const brandId = useProductSelector(productStore.selectBrandId)
const useBrandSelector = createUseSelectorWithParam(brandId)
const subTitle = useBrandSelector(brandStore.selectLabel)
return <Tile title={title} value={value} subTitle={subTitle} />
}
```
We currently use this "higher-order function" (HOF) approach in several thousand places across the codebase, with code like the following:
```jsx
export const createUseSelectorWithParam = (param) => {
const useSelectorWithParam = (selector) =>
useSelector((state) => selector(state, param))
return useSelectorWithParam
}
```
This approach reduces code complexity (fewer arrow functions) and enhances team productivity by decreasing informational noise. However, it currently lacks compatibility with the new React Compiler.
**Question**: Is there a way to inform the React Compiler that the result of `createUseSelectorWithParam` should be treated as a stable hook?
We're hesitant to replace thousands of instances across the codebase with arrow functions or to add parameters universally, as it would likely reduce readability (sometimes only one parameter is needed to access Redux store data, but other times two or even three are required). Additionally, with a higher number of parameters, making such extensive changes could lead to more bugs, as manually adjusting parameters in multiple places increases the chance of errors.
### How often does this bug happen?
Every time
### What version of React are you using?
18.2
### What version of React Compiler are you using?
19.0.0-beta-6fc168f-20241025 | open | ku8ar | 2024-10-30T15:29:52 | 2025-11-30T12:32:51 | null | [
"Type: Feature Request",
"Component: React Compiler"
] | null | 0 | {
"+1": 1
} | 7 | false | 2026-03-29T04:51:56.637439 | |||
facebook | react | 35,260 | I_kwDOAJy2Ks7bXnNm | false | Bug: | <!--
Please provide a clear and concise description of what the bug is. Include
screenshots if needed. Please test using the latest version of the relevant
React packages to make sure your issue has not already been fixed.
-->
React version:
## Steps To Reproduce
1.
2.
<!--
Your bug will get fixed much faster if we can run your code and it doesn't
have dependencies other than React. Issues without reproduction steps or
code examples may be immediately closed as not actionable.
-->
Link to code example:
<!--
Please provide a CodeSandbox (https://codesandbox.io/s/new), a link to a
repository on GitHub, or provide a minimal code example that reproduces the
problem. You may provide a screenshot of the application if you think it is
relevant to your bug report. Here are some tips for providing a minimal
example: https://stackoverflow.com/help/mcve.
-->
## The current behavior
## The expected behavior
| closed | completed | tirth-shah-digipine | 2025-12-01T10:48:55 | 2025-12-01T10:49:25 | 2025-12-01T10:49:25 | [
"Status: Unconfirmed"
] | null | 0 | {} | 0 | false | 2026-03-29T04:51:56.696399 | ||
facebook | react | 35,203 | I_kwDOAJy2Ks7aMM5P | false | [Compiler Bug]: React compiler discards computed property keys when used in method shorthand | ### What kind of issue is this?
- [x] React Compiler core (the JS output is incorrect, or your app works incorrectly after optimization)
- [ ] babel-plugin-react-compiler (build issue installing or using the Babel plugin)
- [ ] eslint-plugin-react-hooks (build issue installing or using the eslint plugin)
- [ ] react-compiler-healthcheck (build issue installing or using the healthcheck script)
### Link to repro
https://playground.react.dev/#N4Igzg9grgTgxgUxALhHCA7MAXABOgWwAcpsEATABRgiIGkEBPXAXlwB0QAzCCAIwCGMTgG52GLlAxxsAS0y4AYrwAUASlzBxuHbhgJssDLgA85WQDcAfNt12TAISG5+AKxbBguANqESZKhp6JgBddU1cAF8o6IB6G2NdE1jza3FIkAAaNEwuWQBzFBBZYggYPGxGIgQIygAbKHzZDAB5IjlMMCjcLhoCXAByQT4EOoBaIgamjDH9ARkxv1k6hBgU2RwBsQxxFS1E2NiluoEOjABZCHIEZA4QATq6znTcMFONvIQu+sbmtrOwGoRFlwAALCAAdwAkhgyDAMA8wCguIiEJEgA
### Repro steps
See the playground link for the full example.
pass an object with a computed key to a react component
eg:
```tsx
<Bar obj={{ [computedPropKey]() { } }} />
```
observe the compiler generates incorrect code
```tsx
<Bar obj={{ computedPropKey() {} }} />
```
### How often does this bug happen?
Every time
### What version of React are you using?
latest on the compiler playground
### What version of React Compiler are you using?
latest on the compiler playground | open | sadan4 | 2025-11-24T21:45:10 | 2025-11-29T20:14:47 | null | [
"Type: Bug",
"Component: React Compiler"
] | null | 0 | {
"heart": 1
} | 1 | false | 2026-03-29T04:51:56.749005 | |||
facebook | react | 34,038 | I_kwDOAJy2Ks7DGC3I | false | Bug: `<dialog>` onClose and onCancel events bubbling | React version: 19.1.1
Chrome version: 138.0.7204.169
## Steps To Reproduce
1. Click "Open Outer Dialog"
2. Click "Open Inner Dialog"
3. Click "Close Inner Dialog" or press Esc on keyboard
Link to code example:
React repro: https://codesandbox.io/p/sandbox/qkndj4
Plain JS comparison: https://codepen.io/heguro/pen/XJmNeJr
## The current behavior
In the React repro, clicking "Close Inner Dialog" produces the following logs:
```
inner dialog: onClose
outer dialog: onClose
div: onClose
```
In plain JS, the [close](https://developer.mozilla.org/en-US/docs/Web/API/HTMLDialogElement/close_event) and [cancel](https://developer.mozilla.org/en-US/docs/Web/API/HTMLDialogElement/cancel_event) events do not bubble, so the same operation only outputs `inner dialog: onClose`.
## The expected behavior
Should only output the following log, same as plain JS behavior:
```
inner dialog: onClose
```
## Additional context
In `packages/react-dom-bindings/src/events/DOMPluginEventSystem.js`, both `'close'` and `'cancel'` are listed in `nonDelegatedEvents`:
https://github.com/facebook/react/blob/33a2bf78c47bf3298655908eec123f5fedbebcc7/packages/react-dom-bindings/src/events/DOMPluginEventSystem.js#L233-L250
However, in `packages/react-dom-bindings/src/events/plugins/SimpleEventPlugin.js`:
https://github.com/facebook/react/blob/33a2bf78c47bf3298655908eec123f5fedbebcc7/packages/react-dom-bindings/src/events/plugins/SimpleEventPlugin.js#L199-L209
The TODO comment mentions plans to add all `nonDelegatedEvents` to prevent bubbling, originally planned for React 18, but we're now at React 19 and `'close'` and `'cancel'` events still bubble incorrectly. (Related PR: #19464 #19761)
Edit:
- `event.stopPropagation()` can be used as a workaround to prevent bubbling.
- Note that the `'cancel'` event [doesn't bubble on dialog](https://developer.mozilla.org/en-US/docs/Web/API/HTMLDialogElement/cancel_event) but [does on input elements](https://developer.mozilla.org/en-US/docs/Web/API/HTMLInputElement/cancel_event) in the DOM. See w3c/webref/issues/1212 | closed | not_planned | heguro | 2025-07-29T10:47:09 | 2025-11-26T21:56:49 | 2025-11-22T18:06:41 | [
"Status: Unconfirmed",
"Resolution: Stale"
] | null | 0 | {} | 4 | false | 2026-03-29T04:51:56.835125 | ||
facebook | react | 27,658 | I_kwDOAJy2Ks519q3y | false | Bug: Seems like new hydrateRoot method doesn't provide any promise or callback to catch hydration process finish | With deprecated hydrate function i could do
`hydrate(
<Component {...properties} />,
container,
resolve
));`
So resolve would be called after hydration finished, but it seems impossible with new hydrateRoot
React version: 1818.2.0
## The expected behavior
`hydrateRoot(
container,
<Component {...properties} />,
{},
resolve
));`
or
`hydrateRoot(
container,
<Component {...properties} />,
{
onHydrationFinished: resolve
}
));`
| closed | not_planned | 123cmp | 2023-11-06T13:02:19 | 2025-12-01T14:58:56 | 2024-04-13T05:04:16 | [
"Status: Unconfirmed",
"Resolution: Stale"
] | null | 0 | {} | 3 | false | 2026-03-29T04:51:56.870024 | ||
facebook | react | 35,261 | I_kwDOAJy2Ks7bY1FV | false | Bug: React 19.2 doesn't seem to have useEffectEvent :( | React 19.2 doesn't seem to have useEffectEvent :(
<img width="830" height="715" alt="Image" src="https://github.com/user-attachments/assets/03a370be-beba-4e1b-a747-67bd73d45ff4" />
React version:
19.2.0
## The current behavior
## The expected behavior
| closed | not_planned | Dorious | 2025-12-01T12:13:57 | 2025-12-01T15:07:57 | 2025-12-01T14:35:21 | [
"Status: Unconfirmed"
] | null | 0 | {} | 4 | false | 2026-03-29T04:51:56.928737 | ||
facebook | react | 31,702 | I_kwDOAJy2Ks6icEHX | false | [Compiler Bug]: server components built externally by the compiler error when rendered in app | ### What kind of issue is this?
- [X] React Compiler core (the JS output is incorrect, or your app works incorrectly after optimization)
- [ ] babel-plugin-react-compiler (build issue installing or using the Babel plugin)
- [ ] eslint-plugin-react-compiler (build issue installing or using the eslint plugin)
- [ ] react-compiler-healthcheck (build issue installing or using the healthcheck script)
### Link to repro
https://github.com/DanielOrtel/compiler-bug
### Repro steps
I'm building an esm UI library and ran into a weird issue with one component when adding react compiler to the build pipeline(which uses rollup with babel). You can check the reproduction repo above, steps:
- set node 20.18.0 if not set automatically
- run `yarn install`
- run `yarn test-compiler-error` // this will build the library and start a bare-bones next app
- open localhost:3000\
- app should error out with the following:
```
TypeError: Cannot read properties of undefined (reading 'H')
at IconsRoot (../../libraries/ui/.dist/icons-root/index.js:12:72)
at AsyncLocalStorage.run (node:async_hooks:346:14)
at stringify (<anonymous>)
at stringify (<anonymous>)
at AsyncLocalStorage.run (node:async_hooks:346:14)
at AsyncResource.runInAsyncScope (node:async_hooks:206:9)
digest: "2235527820"
15 | require("next/dist/compiled/react").__CLIENT_INTERNALS_DO_NOT_USE_OR_WARN_USERS_THEY_CANNOT_UPGRADE;
16 | exports.c = function (size) {
> 17 | var dispatcher = ReactSharedInternals.H; // <-- ReactSharedInternals is undefined here.
```
This is a limited snippet of the larger UI library. The error occurs because the file doesn't have the `'use client'` directive, but this component should be server-renderable, it's just a barebones `div` component. If I remove the `react-compiler` from the build pipeline or if I make this a client component, it works as expected.
Ultimately, I'd expect adding the compiler to the build pipeline of my library to not break components which previously could be server rendered and force the use of client directive when it is unnecessary.
My suggestion, in case compiler-optimized components can't be rendered in server components, would be to have some config option that can be enabled for libraries to check for client directives and only optimize client components. And also to add some form of an error boundary to let consumers know that they're trying to render a compiler-optimized component in a server environment. The current error is just confusing and it took me a while to figure out what was the exact cause of it.
But the ideal scenario would be to not have to worry about in which environment the component is rendered. Sure, the compiler optimizations do nothing on the server, but that means that they could be essentially stubbed and still render as expected. I dislike the idea of having to output separate versions of components that can be potentially server rendered, simply because I'd like the client versions to be optimized by the compiler.
### How often does this bug happen?
Every time
### What version of React are you using?
19
### What version of React Compiler are you using?
19.0.0-beta-df7b47d-20241124 | open | DanielOrtel | 2024-12-08T14:09:14 | 2025-11-27T14:47:40 | null | [
"Type: Bug",
"Component: React Compiler"
] | [
"eps1lon"
] | 0 | {
"+1": 5
} | 16 | false | 2026-03-29T04:51:56.966327 | |||
facebook | react | 34,259 | I_kwDOAJy2Ks7HR68c | false | Bug: in dev mode the dispose function is applied to incorrect state | When using a third-party stateful object that needs to be explicitly disposed of, I'm using an effect that returns a dispose function. In dev mode I observe then that React uses the stateful object after it has been disposed, and then disposes of it a second time.
React version: 19.1.1
## Steps To Reproduce
1. Run the code sample below
2. Observe the console logs
3. Navigate away to some other page
```ts
import React from "react";
export const TestComponent = () => {
console.log("Start rendering TestComponent");
const state = React.useMemo(() => {
return new State();
}, []);
React.useEffect(() => {
return () => state.dispose();
}, []);
console.log("TestComponent rendered", state.id);
return "Test";
};
class State {
id: number;
constructor() {
this.id = Math.floor(Math.random() * 1000000);
console.log("new State", this.id);
}
dispose() {
console.log("dispose of state", this.id);
}
}
```
## The current behavior
Currently, the component is created twice in dev mode (as expected), and each copy creates a new State (as expected). One of those copies then gets disposed during unmount (as expected). However, that same object gets used again in a subsequent render even after it was unmounted (not expected). It then gets disposed of again during subsequent navigation (definitely not expected). The original second copy of the State is not disposed (also not expected).
```
Start rendering TestComponent
new State 666058
new State 109916
TestComponent rendered 666058
Start rendering TestComponent
TestComponent rendered 666058
dispose of state 666058
Start rendering TestComponent
TestComponent rendered 666058
Start rendering TestComponent
TestComponent rendered 666058
(navigate)
dispose of state 666058
```
## The expected behavior
```
Start rendering TestComponent
new State 666058
new State 109916
TestComponent rendered 666058
Start rendering TestComponent
TestComponent rendered 109916
dispose of state 109916
Start rendering TestComponent
TestComponent rendered 666058
Start rendering TestComponent
TestComponent rendered 666058
(navigate)
dispose of state 666058
``` | closed | not_planned | pasha-omni | 2025-08-21T22:18:08 | 2025-12-01T17:07:16 | 2025-12-01T17:07:16 | [
"Status: Unconfirmed",
"Resolution: Stale"
] | null | 0 | {} | 2 | false | 2026-03-29T04:51:57.043523 | ||
facebook | react | 34,286 | I_kwDOAJy2Ks7HuIJY | false | I attempted to use react - devtools, but it failed. | My scenario is that a React component is rendered in an iframe. I hoped to use react - devtools to detect the internal state, but it didn't work out.
OS: macOS 14.7.3
Browser: chrome 139.0.7258.68
1. (my code)
if (window.parent !== window) {
const script = document.createElement("script");
script.src = "http://localhost:8097";
document.head.appendChild(script);
}
2. (console)
npx react-devtools
<img width="1600" height="1200" alt="Image" src="https://github.com/user-attachments/assets/b284ea17-875e-4ea6-8fd2-7c5e8702a176" /> | closed | not_planned | SunJiashun | 2025-08-25T07:25:27 | 2025-12-01T17:07:09 | 2025-12-01T17:07:09 | [
"Resolution: Stale"
] | null | 0 | {} | 3 | false | 2026-03-29T04:51:57.087892 | ||
facebook | react | 33,197 | I_kwDOAJy2Ks62kK5c | false | Bug: Conditional button rendering with button and submit inside a form causes type="button" to submit | React version: 19.1.0
## Steps To Reproduce
1. Render a `<form>` that uses `onSubmit`.
2. Inside it, conditionally render a button based on state (in the example, `step`).
3. Initially, render a `type="button"` with an `onClick` handler (`nextFunc()`).
4. When clicked, it updates the state.
5. But when clicked, both `onClick` (`nextFunc`) and the form's `onSubmit` (`submitFunc`) are fired, even though the button is `type="button"` and not `type="submit"`.
Link to code example: https://codesandbox.io/p/sandbox/strange-murdock-n82pjy (open the console to check the logs)
## The current behavior
When clicking the "Next" button (`type="button"`), BOTH the button’s `onClick` and the form’s `onSubmit` handlers are triggered.
_Important Note_: **This behavior does not occur if each button is given a unique key when rendered.**
## The expected behavior
The `type="button"` button should trigger ONLY its `onClick` handler and the `type="submit"` button should trigger ONLY the `onSubmit` handler. | closed | not_planned | xCirno1 | 2025-05-14T12:23:58 | 2025-12-01T17:07:38 | 2025-12-01T17:07:38 | [
"Status: Unconfirmed",
"Resolution: Stale"
] | null | 0 | {
"eyes": 2,
"heart": 1
} | 6 | false | 2026-03-29T04:51:57.165618 | ||
facebook | react | 34,285 | I_kwDOAJy2Ks7Htis0 | false | [Proposal] Align suspense behavior | ## Summary
I propose that every `react` root be automatically wrapped in a `Suspense` boundary with no fallback
UI.
```diff
import { createRoot } from 'react-dom/client';
import { App } from './app';
const element = document.getElementById('root');
const root = createRoot(element);
root.render(
+ // Suspense boundary with no fallback UI added by react internally
+ <Suspense fallback={null}>
<App />
+ </Suspense>
);
```
## Details
In order to understand why I am proposing this change, let's take a look at the suspense behavior in
`react@19.1.1` and `react-dom@19.1.1`.
I've created an application where one component is a lazy component:
```tsx
import { lazy } from 'react';
const LazyContent = lazy(() => import('./content'));
```
### Scenario 1: Lazy component in initial tree
In this scenario, a `lazy` component is rendered as a part of the initial `react` tree. The `lazy`
component will always take longer than `2000ms` to resolve.
#### Case 1.1: With Suspense boundary
[Standalone example](https://stackblitz.com/edit/vitejs-vite-tuduj4gz?file=src%2Fapp.tsx)
| Initial render | After dynamic import resolved |
| -------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------ |
|  |  |
| `Suspense` fallback UI rendered | Successful tree render |
#### Case 1.2: No Suspense boundary
[Standalone example](https://stackblitz.com/edit/vitejs-vite-ezymz8y5?file=src%2Fapp.tsx)
| Initial render | After dynamic import resolved |
| -------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------ |
|  |  |
| `Suspense` fallback UI rendered | Successful tree render |
Having a "blank" fallback UI makes sense to me for the initial render. The `lazy` component has
thrown a promise (with the dynamic import) and `react` does not have a tree to render. So `react`
renders nothing.
### Scenario 2: Conditional lazy component
In this scenario, a `lazy` component is not initially in the `react` tree, and is added into the
tree after clicking the "Add lazy content" button. The `lazy` component will always take longer than
`2000ms` to resolve.
#### Case 2.1: With Suspense boundary
[Standalone example](https://stackblitz.com/edit/vitejs-vite-hirl1dfu?file=src%2Fapp.tsx)
| Before loading | During loading | Loaded |
| -------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------ |
|  |  |  |
| Lazy component not yet in `react` tree | Lazy component in the `react` tree and `Suspense` fallback UI is rendered | Lazy loaded content rendered |
#### Case 2.2: No Suspense boundary
[Standalone example](https://stackblitz.com/edit/vitejs-vite-uk25rswn?file=src%2Fapp.tsx)
| Before loading | During loading | Loaded |
| -------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------ |
|  |  |  |
| Lazy component not yet in `react` tree | Existing UI preserved | Lazy loaded content rendered |
### Problems
There are a few problems with the current way `react` handles components that throw promises (eg
`lazy`) when there is no `Suspense` boundary.
#### 1. Behaviour mismatch
There is a mismatch in behaviour between an initial render of a `react` root with `lazy` components
(Case 1.2), and later conditional rendering of `lazy` components (Case 2.2).
- **Initial render**: behaves as if there is a `Suspense` with no fallback UI at the root
- **Later conditional renders**: behaves as if there is a `Suspense` boundary around the `lazy`
component. Well, sort of. When there is no `Suspense` boundary, the existing UI is _preserved_
which is different to when you use a `Suspense` boundary where the fallback UI is rendered.
If a `Suspense` boundary were automatically added around the root, then there would be no mismatch
between initial renders and later conditional renders (both cases use an empty fallback UI).
_Wouldn't showing a blank screen worse than the current state?_
- The mismatch in how `lazy` is handled can lead to confusion. Folks might think that their
application is looking good (with a conditional `lazy` component and no `Suspense` boundary) and
when they start using their `lazy` component in an initial render, the suspending behaviour of
their application is different, which I expect would be confusing.
- Showing a blank screen would make it clearer that the application is suspending, and encourage
people to always explicitly handle loading states (see below)
#### 2. Adding a `Suspense` boundary can lead to surprises
- Let's say that you are using `lazy` without a `Suspense` boundary, and your application is leaning
on the existing UI being preserved until the `lazy` component is ready.
- At a later point, either yourself or somebody else comes in and adds a `Suspense` boundary higher
in your `react` tree.
- All of a sudden the experience of your component changes (the higher `Suspense` boundary now
starts showing the fallback UI rather than the existing "pausing" UI behaviour), and it might not
be obvious why this is happening to people.
If there was always a `Suspense` boundary at the root, all suspense behaviour would be handled in
the same way from the start, leading to less surprises.
#### 3. No loading states for users
- The current behaviour leads to experiences where loading states are not explicitly handled. People
use `lazy` without `Suspense` and it will work, but there is no loading state (unless the maker
adds their own loading state on top of suspense).
- This problem might be invisible to folks, as on their local machine the `lazy` might resolve
quickly, whereas in production the `lazy` might take some time to complete.
Using a `Suspense` boundary with no fallback at the root would be a forcing function for makers to
need to handle loading states for suspending UI - otherwise they will get an empty root while the
root is suspended, which makes it clearer that loading states need to be handled.
### Development warnings
In addition to adding `<Suspense fallback={null}>` around a `react` root, I think it could be
helpful to also log a development only warning to the `console` when the default `Suspense` boundary
is hit. This will make it easier for makers to know what is happening (that the application is being
suspended and no fallback UI has been provided) and what they should do about it.
### Proposal screenshots
Here are the screenshots of what I am proposing to help with clarity:
#### Case 1.2: No Suspense boundary
_Current `react` behavior is unchanged_
| Initial render | After dynamic import resolved |
| -------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------ |
|  |  |
| `Suspense` fallback UI rendered | Successful tree render |
#### Case 2.2: No Suspense boundary
_UI no longer preserved while suspending_
| Before loading | During loading | Loaded |
| -------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------ |
|  |  |  |
| Lazy component not yet in `react` tree | No fallback UI | Lazy loaded content rendered |
### Alternatives
Rather than adding a Suspense boundary around a root with an empty fallback UI, there could be other
ways to handle this situation. For example, `react` could throw an error when a root suspends and
there is no `Suspense` boundary. Personally, I think having `<Suspense fallback={null}>` around a
`react` root and a development warning would be a nicer path.
### Versioning
If this proposal is accepted, it would likely need to be shipped in a `major` version.
---
Thank you for your consideration and for your amazing ongoing work on `react`,
Cheers | closed | not_planned | alexreardon | 2025-08-25T06:25:33 | 2025-12-01T17:07:11 | 2025-12-01T17:07:10 | [
"Status: Unconfirmed",
"Resolution: Stale"
] | null | 0 | {
"heart": 1,
"rocket": 2
} | 4 | false | 2026-03-29T04:51:57.275770 | ||
facebook | react | 34,982 | I_kwDOAJy2Ks7T0Mv6 | false | Bug: | I found that in the react render, the class `SRu4RHny` and `BIxkyKps` just like here below:
in the pure html file, the `SRu4RHny` will be rendered first and then render the `BIxkyKps` class props,
but why the `SRu4RHny` will be rendered in the last ?
<img width="1049" height="588" alt="Image" src="https://github.com/user-attachments/assets/56e52425-e63e-4506-98bf-f97f755bfdb6" />
<img width="888" height="551" alt="Image" src="https://github.com/user-attachments/assets/2b39fe7f-f6aa-47bf-a4e5-719238a5e974" />
## The current behavior
SRu4RHny rendered in the last and the BIxkyKps was rendered in the fisrt
## The expected behavior
just like the pure html, the `BIxkyKps` should be rendered in the last , as it was placed in the behind, just like the pure html
| closed | completed | BertramYe | 2025-10-26T09:51:01 | 2025-12-02T09:42:52 | 2025-12-02T09:42:52 | [
"Status: Unconfirmed"
] | null | 0 | {} | 1 | false | 2026-03-29T04:51:57.355230 | ||
facebook | react | 35,024 | I_kwDOAJy2Ks7VC8YM | false | [Compiler Bug]: TanStack Form incompatibility | ### What kind of issue is this?
- [x] React Compiler core (the JS output is incorrect, or your app works incorrectly after optimization)
- [ ] babel-plugin-react-compiler (build issue installing or using the Babel plugin)
- [ ] eslint-plugin-react-compiler (build issue installing or using the eslint plugin)
- [ ] react-compiler-healthcheck (build issue installing or using the healthcheck script)
### Link to repro
https://playground.react.dev/#N4Igzg9grgTgxgUxALhASwLYAcIwC4AEwBUYCAYmggDYAmAwhAHZ4IAehAvgQGYwQYCAcgB0Aeh64MAWjjNWHEXjBshAHSYb2OfLyhM4eNMwIAVdnko1aAOQh4A6rgDWaJgHMAFMWoBDAEY0ADQEABZotLQITAScyEQEfoHU8WB4MG7uIeGR0fH+EBDUCL4xnACURBoEBHJMabxUdAQAvCRkVnSMLBYAPGkZHgB8nuXVBDAIeLAxnuM1vUk0Q-M1BP1YpUPAS9ScvWJgm0wrmnhq53jAAIQ5UTEAZA8aF6+XF71uWFDnZ2--eAAbr5qFAEC1gDwmrQRGlfKwRMDQQhOC93uiLsx6KFSu5wcBPAhKi0ho1rCIcUxaMVsbiEISlL4YHi8IiQWDyqi-hi8GJTgCuWt1mJdqcamMmFytGwdIQePpDMYYuYOJ1aE4YK4PN5EgEaLF4j49SkCANMrFKsA0XUGlDrK12hRod0FIReqb0ubSaM0ZNpjBZmiPqKg5cNlsdsb9odjvz0Z8mN9fgC3kiwRC7XRYXh4Qg2cipTzMUxaR58YTiaTMzDKdSEKW8Qyc8ypvmOYWAXzQwcQ2cJZwQEEQHUoe4UOhsLhCHgAJ5YBAJAAKoPcbgA8lgjMwwLFePxBEJ-MbpFgV25pJNfIZZAIsGhijAxLQ0GkhABuF5MbzjMRiOTYe94SVABZCAoniNQQBBahII0bgwCAsA7R3ZcoFXJgNy3epyjfQdwFCCAAHcAEkegDEEwBQHgKJRIA
### Repro steps
When rendering something with an expression conditionally like this `{!hidden && <input value={field.state.value} onChange={(e) => field.handleChange(e.target.value)}/>}`, the "memo dependencies" to trigger a rerender are hidden and field.
This is different when rendering the input without a condition like this `{<input value={field.state.value} onChange={(e) => field.handleChange(e.target.value)}/>}`, the "memo dependencies" are "field.handleChange" and "field.state.value".
This change of behavior can cause some unexpected issues:
- in my case, using @tanstack/form, the field property isn't reactive in itself but its subproperties are. This prevents the input from updated its value when the user types inside the input
- if field was reactive, a change inside field that is neither the value or handleChange would trigger a rerender that shouldn't be needed
This can happen with a `bool && <div>` or with a `bool ? <div> : <div>` expression
Not really sure if this is expected behavior and if it is, what is the recommended fix.
I've opened an [issue](https://github.com/TanStack/form/issues/1832) over at @tanstack/form
### How often does this bug happen?
Every time
### What version of React are you using?
19.0.0
### What version of React Compiler are you using?
1.0.0 | open | JamesHyot | 2025-10-31T10:21:45 | 2025-12-02T19:30:31 | null | [
"Type: Bug",
"Status: Unconfirmed",
"Component: React Compiler"
] | null | 0 | {
"+1": 1,
"eyes": 1,
"heart": 1
} | 4 | false | 2026-03-29T04:51:57.394700 | |||
facebook | react | 35,272 | I_kwDOAJy2Ks7b2Wmf | false | Bug: | <!--
Please provide a clear and concise description of what the bug is. Include
screenshots if needed. Please test using the latest version of the relevant
React packages to make sure your issue has not already been fixed.
-->
React version:
## Steps To Reproduce
1.
2.
<!--
Your bug will get fixed much faster if we can run your code and it doesn't
have dependencies other than React. Issues without reproduction steps or
code examples may be immediately closed as not actionable.
-->
Link to code example:
<!--
Please provide a CodeSandbox (https://codesandbox.io/s/new), a link to a
repository on GitHub, or provide a minimal code example that reproduces the
problem. You may provide a screenshot of the application if you think it is
relevant to your bug report. Here are some tips for providing a minimal
example: https://stackoverflow.com/help/mcve.
-->
## The current behavior
## The expected behavior | closed | not_planned | Vchenrd | 2025-12-03T04:22:03 | 2025-12-04T05:13:11 | 2025-12-04T05:13:11 | [
"Status: Unconfirmed"
] | null | 0 | {} | 0 | false | 2026-03-29T04:51:57.444005 | ||
facebook | react | 34,752 | I_kwDOAJy2Ks7P2JQX | false | [Compiler Bug]: Compiler incorrectly assumes non-nullability and lifts property access | ### What kind of issue is this?
- [x] React Compiler core (the JS output is incorrect, or your app works incorrectly after optimization)
- [ ] babel-plugin-react-compiler (build issue installing or using the Babel plugin)
- [ ] eslint-plugin-react-compiler (build issue installing or using the eslint plugin)
- [ ] react-compiler-healthcheck (build issue installing or using the healthcheck script)
### Link to repro
https://playground.react.dev/#N4Igzg9grgTgxgUxALhHCA7MAXABAUQA8BDAWwAcAbBXAXlwAoBKOgPl2AB0Ndd0s8AbRgIw5TGAQAaXJOwAlUeKwIAunVxRJAZWzFsCADzBcAE33FkuDFEqVcAX1wAfDmYtWTAMwQJTAI2I4AGsrHBgASwwAcwBuXGwAT3IEK05wKDhEMDB0x3zXE3M9TwTk1Nx0hBgYCBg8pwdWBiKPa1t7ByZY7m5efhxcAAtiDFNqbUzssC9bSf9SCJyIzA0tBHwvHzhsfAA3BAxsBmY2Dj7eXGIc6uORMQkEADpi4gB+J6SUulp6dLApqJciAZOkACpDJayIbQSimXD+GhRPYQYJ+XCYSiJXAAdyGh3celwUIBWSBszs2IBCwi2AMpnS3V6PEuAwg1CelAg0RauB8fkCISs92UkheFie-ICQWCjiZLK6PQwF3Wm22x1OtHYXBZvAiXkYIse4r0Hy+NF+fwyZJyjPOusuIzGE0BOQp80Wy0wzCVl0cFwcMmESmNr1U8ouImwsB4DAuvEM-igdMw8cumAAwpQIiFaMBNdq0365IoHipea9SlLBaFKiADDh0jJzWlrdMGnLfX7HA4i2UUrR0kmU8qQGnWH2ABIIOwQXF1OFpwwAemH2EwE5Z8ocSpBaEwXgi0RQIAiFDqeHNbgACpQoNEogB5cjYFZYfJeWqkXAAckCiMoABaKh7yiQCRCCbBAPQCgImoGBl1MJZsB-JVuBaC5l2XGDyDg-Q3wAWQgUwKnSYg7HSbgnDAfCZgiURcFvUCMGfV8JG6PcwBhHEAEkjmqDByLAFAvCEhAHCAA
### Repro steps
I'm using React Router fetchers for mutations, while migrating the app to 19.2.0 and trying to migrate some `useRef()` latest pattern to `useEffectEvent` I spotted weird bug with something being accessed despite being undefined at render.
Basically, having a function that's supposed to be invoked only when some state is in shape X (e.g. successful response) eagerly accesses state that this function uses to cache it. It causes the error "cannot read property of undefined" error to occur
<img width="924" height="735" alt="Image" src="https://github.com/user-attachments/assets/92522be2-eb1d-47cc-a394-25f2836c6372" />
1. I'm having a state which is an union of few different states (in original example this is RR `useFetcher`)
2. I'm having a `useEffectEvent` callback that's supposed to be called within `useEffect` when data becomes successful, additionaly I have an assertion so TypeScript allows me to access the `.feedback` part and no-one tries to invoke the callback otherwise
3. Mentioned `useEffect` that invokes (2)
4. User action invokes state change (in original example this is asynchronous, and happens in background, hence `useEffect` usage at all).
In the output the data is eagerly accessed to cache the function, causing the error:
<img width="767" height="416" alt="Image" src="https://github.com/user-attachments/assets/51949118-d109-47b4-a2f8-95ab7b31701b" />
Random thoughts: maybe it's somehow partially related to my `assert` function, where compiler doesn't know what does it do? But it causes me to don't use optional chaining, but compiler doesn't know why I'm not using it and is not informed by the types 🤔 Also it's not strictly related to the `useEffectEvent` or anything like that, because if I remove it - it'll still cause the same problem
### How often does this bug happen?
Every time
### What version of React are you using?
19.2.0
### What version of React Compiler are you using?
19.1.0-rc.3 | open | pawelblaszczyk5 | 2025-10-06T12:36:38 | 2025-12-02T03:10:46 | null | [
"Type: Bug",
"Status: Unconfirmed",
"Component: React Compiler"
] | null | 0 | {
"+1": 11,
"heart": 1
} | 15 | false | 2026-03-29T04:51:57.532701 | |||
facebook | react | 35,273 | I_kwDOAJy2Ks7b2XVg | false | Bug: | <!--
Please provide a clear and concise description of what the bug is. Include
screenshots if needed. Please test using the latest version of the relevant
React packages to make sure your issue has not already been fixed.
-->
React version:
## Steps To Reproduce
1.
2.
<!--
Your bug will get fixed much faster if we can run your code and it doesn't
have dependencies other than React. Issues without reproduction steps or
code examples may be immediately closed as not actionable.
-->
Link to code example:
<!--
Please provide a CodeSandbox (https://codesandbox.io/s/new), a link to a
repository on GitHub, or provide a minimal code example that reproduces the
problem. You may provide a screenshot of the application if you think it is
relevant to your bug report. Here are some tips for providing a minimal
example: https://stackoverflow.com/help/mcve.
-->
## The current behavior
## The expected behavior | closed | not_planned | Vchenrd | 2025-12-03T04:23:28 | 2025-12-04T05:13:15 | 2025-12-04T05:13:15 | [
"Status: Unconfirmed"
] | null | 0 | {} | 0 | false | 2026-03-29T04:51:57.624193 | ||
facebook | react | 35,231 | I_kwDOAJy2Ks7avfjd | false | Bug: Build of `react-dom` hangs on Deno runtime |
React version: 19.2.0
## Steps To Reproduce
1. clone https://github.com/Yovach/deno-tanstack-start-bug-reproduction
2. deno install
3. deno task build
## The current behavior
It's hanging forever.
## The expected behavior
Not hanging on `deno task build`
## Context
I've been investigating the vite build hanging issue in Deno and found the root cause. https://github.com/denoland/deno/issues/31248
Seems that `react-dom` package.json maps Deno to use the browser bundle (`server.browser.js`) instead of the node bundle. The browser bundle creates a MessageChannel for task scheduling, and this MessageChannel keeps Deno's process alive forever, causing the hang.
Runtime mappings:
• Node.js → uses `setImmediate` from server.node.js (doesn't block exit) :white_check_mark:
• Bun → uses `queueMicrotask` from server.bun.js (doesn't block exit) :white_check_mark:
• Deno → uses `MessageChannel` from server.browser.js (blocks exit forever) :x:
## Solutions
Changing the "deno" mapping from `"./server.browser.js"` to `"./server.node.js"` at `node_modules/.deno/react-dom@19.2.0/node_modules/react-dom/package.json` fixed the hang.
I also tested changing the bun bundler to use server.browser.js, to have sure that this is the problem, and it hangs for bun too.
I talked with the Deno CLI team, and we decided to create this issue to check if make sense have a new bundle for Deno. For some reason, seems that solution was rejected in the past (https://github.com/facebook/react/pull/30655#issuecomment-2283099574)
If it does not make sense again, we will need to add a special condition for `react-dom` on Deno side.
| closed | completed | fraidev | 2025-11-27T05:08:45 | 2025-12-04T06:02:50 | 2025-12-04T05:50:28 | [
"Status: Unconfirmed"
] | null | 0 | {} | 1 | false | 2026-03-29T04:51:57.657317 | ||
facebook | react | 35,268 | I_kwDOAJy2Ks7brqRt | false | [Compiler Bug]: incorrectly uses variable name as JSX tag (`<base />`) instead of referenced value | ### What kind of issue is this?
- [x] React Compiler core (the JS output is incorrect, or your app works incorrectly after optimization)
- [ ] babel-plugin-react-compiler (build issue installing or using the Babel plugin)
- [ ] eslint-plugin-react-hooks (build issue installing or using the eslint plugin)
- [ ] react-compiler-healthcheck (build issue installing or using the healthcheck script)
### Link to repro
https://playground.react.dev/#N4Igzg9grgTgxgUxALhASwLYAcIwC4AEASggIZyEBmMEGBA5DGRfQNwA6AdnBJ2IQCNSYBAQC8DACZoAbmy5cefQgBUE-AMK0cnBJzzJizPADoAYhvEEAFAEpxAPgLAuBAkv4Et2K0JEdONyY8WECAHm8sAHoHLgBfAK4EAA8cfAJJBEpSKAAbKihuPDReAgAJWgQ7Z1cCYNCCMLVNbV49Qhj4rhAAGhAlSjQAcxR0bFxCPABPLFFgAgAFXKghtE4AeSxi3jACOIJqWgYhAQRcgFosZdXOc6ZyPHOebDRchBgo6X55Ti5rF0CBCiUWeWFepG2nAAshBMoZ2CBSLlcgj4gQwBC0GBBupFtc1ptIWBbKxeuAABYQADuAEl9O9OEiwChsrkRHEgA
### Repro steps
Given this input:
```jsx
import React from 'react';
const base = 'div';
const TestComponent: React.FC = () => {
const Comp = base;
return <Comp/>
};
export default function Home() {
return <TestComponent />
}
```
The compiler outputs this for the `TestComponent`
```jsx
const TestComponent: React.FC = () => {
const $ = _c(1);
let t0;
if ($[0] === Symbol.for("react.memo_cache_sentinel")) {
t0 = <base />;
$[0] = t0;
} else {
t0 = $[0];
}
return t0;
};
```
First reported here: https://github.com/vercel/next.js/issues/86728
This outputs a `<base />` tag instead of a `div`.
I searched on issues before posting, but I couldn't find other reports.
### How often does this bug happen?
Every time
### What version of React are you using?
Next.js' version on 16.0.6
### What version of React Compiler are you using?
1.0.0 | open | icyJoseph | 2025-12-02T12:42:48 | 2025-12-03T07:19:39 | null | [
"Type: Bug",
"Status: Unconfirmed"
] | null | 0 | {} | 2 | false | 2026-03-29T04:51:57.745712 | |||
facebook | react | 35,218 | I_kwDOAJy2Ks7ai_SG | false | [Compiler Bug]: The library for web and native user interface | ### What kind of issue is this?
- [x] React Compiler core (the JS output is incorrect, or your app works incorrectly after optimization)
- [ ] babel-plugin-react-compiler (build issue installing or using the Babel plugin)
- [ ] eslint-plugin-react-hooks (build issue installing or using the eslint plugin)
- [ ] react-compiler-healthcheck (build issue installing or using the healthcheck script)
### Link to repro
Description: React is a JavaScript library for building user interfaces, designed to help developers create interactive, component-based user interfaces for web (and native) applications.
### Repro steps
Description: React is a JavaScript library for building user interfaces, designed to help developers create interactive, component-based user interfaces for web (and native) applications.
### How often does this bug happen?
Every time
### What version of React are you using?
Description: React is a JavaScript library for building user interfaces, designed to help developers create interactive, component-based user interfaces for web (and native) applications.
### What version of React Compiler are you using?
Description: React is a JavaScript library for building user interfaces, designed to help developers create interactive, component-based user interfaces for web (and native) applications. | closed | completed | pinnintipavankumar26-dot | 2025-11-26T09:14:27 | 2025-12-05T16:06:29 | 2025-12-05T16:06:29 | [
"Type: Bug",
"Status: Unconfirmed"
] | null | 0 | {
"eyes": 1
} | 0 | false | 2026-03-29T04:51:57.781408 | ||
facebook | react | 31,359 | I_kwDOAJy2Ks6b4shF | false | [Compiler Bug]: `===` is not a sufficient equality check | ### What kind of issue is this?
- [X] React Compiler core (the JS output is incorrect, or your app works incorrectly after optimization)
- [ ] babel-plugin-react-compiler (build issue installing or using the Babel plugin)
- [ ] eslint-plugin-react-compiler (build issue installing or using the eslint plugin)
- [ ] react-compiler-healthcheck (build issue installing or using the healthcheck script)
### Link to repro
https://playground.react.dev/#N4Igzg9grgTgxgUxALhASwLYAcIwC4AEwBUYCAyngIZ4IEC+BAZjBBgQDogwJVx5cA3BwB2ohAA8c+AgBMETKlAA2hJlBH80EEQQBaCVgAoAlEVEECcHWEIBtAF6GIAGgJk8B1gF0CAXhIyShoEIwAGE2FdKxtCWxh-AgB5ACMAKwR+ADo0MCMnVjcIggB+ThAwrgJkcoBaSpAoyx48WF0AHhSoPDwdAh0AYWU0OABrP2BTfwA+dwRPZyNagogTemngePp2gHounp1pqPpREHogA
### Repro steps
1. Click the rendered button repeatedly
**Expected result:**
On each click, the text should toggle between `0` and `-0`.
**Actual result:**
Initially, the text `0` is rendered. After the first click, `-0` is rendered. Subsequent clicks don't do anything.
### How often does this bug happen?
Every time
### What version of React are you using?
18.3.1
### What version of React Compiler are you using?
19.0.0-beta-8a03594-20241020
---
React Compiler currently uses the `===` operator to check whether to memoize. The `===` operator isn't sufficient to check whether two values are identical. The two edge cases missed by `===` but handled handled by `Object.is()`:
- `0`/`-0`: React Compiler being unable to detect a difference between these values causes this bug.
- `NaN`: React Compiler incorrectly _always_ detects a difference despite one not existing, causes code to not be memoized, but is otherwise correct.
I imagine that changing `===` to `Object.is()` in the output might be bad for performance. In case this bug is a _wontfix_, you should document this limitation somewhere.
| closed | not_planned | Skalman | 2024-10-25T23:10:35 | 2025-12-04T20:19:40 | 2025-06-14T21:04:51 | [
"Component: React Compiler",
"Resolution: Stale"
] | null | 0 | {} | 8 | false | 2026-03-29T04:51:57.831665 | ||
facebook | react | 35,205 | I_kwDOAJy2Ks7aOVkn | false | [Compiler Bug]: Post Increment Assignment Compiling to Pre Increment Assignment | ### What kind of issue is this?
- [x] React Compiler core (the JS output is incorrect, or your app works incorrectly after optimization)
- [ ] babel-plugin-react-compiler (build issue installing or using the Babel plugin)
- [ ] eslint-plugin-react-hooks (build issue installing or using the eslint plugin)
- [ ] react-compiler-healthcheck (build issue installing or using the healthcheck script)
### Link to repro
https://github.com/kbitgood/rc-bug-repro-increment-assign
### Repro steps
#### Description
When using the post increment operator in an assignment (`y = x++`) the value is incremented before being assigned when using the React Compiler.
#### Steps to Reproduce
1. Check out the reproduction repository
2. Run `npm run dev`, or `npm run build && npm run start` as the problem exists in both dev mode and production
3. See the index field of each item in the next list is off by one. Every item in the list should have a sequential "Count" value starting at 1, not
#### Details in the Reproduction Code
The code in `src/app/page.tsx` uses Array.reduce to build an array of nodes, and keeps track of a counter in the aggregated value.
When we use this code:
```ts
const count = agg.itemCounter++;
```
The compiled output translates to:
```ts
agg.itemCounter = agg.itemCounter + 1;
const count = agg.itemCounter;
```
Which is not equivalent because it increments before assigning, rather than assigning and then incrementing.
The issue goes away by taking the counter out of the aggregate object and into a normal variable outside of the reduce function.
#### Conclusion
I realize that this way of keeping the count in the aggregated value is not the best way, but this is not strictly against the "Rules of React". If it is, it should probably be added to the ESLint rules.
This issue was found on a large codebase that was working just fine before enabling the compiler. And we had no indication that this code would not work after enabling the compiler.
### How often does this bug happen?
Every time
### What version of React are you using?
19.2.0
### What version of React Compiler are you using?
1.0.0 | open | kbitgood | 2025-11-25T01:54:26 | 2025-12-05T10:15:41 | null | [
"Type: Bug",
"Component: React Compiler"
] | null | 0 | {
"heart": 2
} | 2 | false | 2026-03-29T04:51:57.962842 | |||
facebook | react | 35,126 | I_kwDOAJy2Ks7Xyn_l | false | Bug: Dev Mode breaks applications: calling getters in props objects causes side-effects | In dev mode, there is a feature that inspects the properties of all props passed to a component, recursively. This happens using a simple `for (key in obj)` loop and thus traverses all enumerable properties, recursively, including getters. This feature cannot be disabled without disabling dev mode and it will only trigger in slightly advanced setups and thus is hard to debug.
Iterating all enumerable keys for getters in an object is a problem, because these getters are also evaluated using simple unconditional `obj[key]` syntax, causing possible side-effects to happen for non-trivial getters.
Although frowned-upon in many cases, it is valid for getters to have side-effects that could change the state of the objects or throw exceptions. Specifically in objects that are unrelated to the React components.
However, in dev mode, this causes the application to break or to behave differently in non-dev mode with hard to diagnose bugs.
React version: 19.2.0
(this is a regression)
## Steps To Reproduce
1. Run the following demo in dev mode
```
import {useEffect, useState} from 'react'
import {createRoot} from "react-dom/client";
function getData() {
const result = {}
result._someInnocentUnusedProp = {}
Object.defineProperty(result._someInnocentUnusedProp, "nestedUnusedCalculatedProp", {
get: () => {
console.log('get nestedUnusedCalculatedProp called - why?')
throw new Error("I should not call getters unconditionally")
}, enumerable: true
})
return result._someInnocentUnusedProp
}
function DataRenderer() { return null }
function Child(props) {
return <div>child: {props.count}<DataRenderer data={getData()}></DataRenderer></div>;
}
function App() {
const [count, setCount] = useState(0);
// just to cause a re-render
useEffect(() => {
if (count > 5) return
setCount((c) => c + 1);
}, [count]);
return <Child count={count}/>;
}
createRoot(document.getElementById("root")).render(<App/>);
```
2. You will get a stacktrace like the following:
```
Uncaught Error: I should not call getters unconditionally
at Object.get [as nestedUnusedCalculatedProp] (main.jsx:10:19)
at addObjectDiffToProperties (react-dom-client.development.js:3968:21)
at addObjectDiffToProperties (react-dom-client.development.js:4016:23)
at logComponentRender (react-dom-client.development.js:4130:22)
at commitPassiveMountOnFiber (react-dom-client.development.js:15469:13)
at recursivelyTraversePassiveMountEffects (react-dom-client.development.js:15439:11)
at commitPassiveMountOnFiber (react-dom-client.development.js:15718:11)
at recursivelyTraversePassiveMountEffects (react-dom-client.development.js:15439:11)
at commitPassiveMountOnFiber (react-dom-client.development.js:15476:11)
at recursivelyTraversePassiveMountEffects (react-dom-client.development.js:15439:11)
```
Link to code example:
[Code Sandbox](https://codesandbox.io/p/sandbox/optimistic-ioana-ct7j24)
(Sometimes you need to reload the page a couple of times for the issue to appear. The timing is not deterministic, which makes the issue especially hard to diagnose.)
## The current behavior
React Dev mode evaluates getters recursively from time to time, possibly causing side effects.
## The expected behavior
Rect dev mode should not evaluate all getters in props objects unconditionally.
## My thoughts
The problematic code is here, which evaluates all enumerable properties, even from prototypes:
https://github.com/facebook/react/blob/3a495ae72264c46b4a4355904c6b4958b0a2f9b2/packages/shared/ReactPerformanceTrackProperties.js#L295
The workaround here (other than not using dev mode) is to make the properties non-enumerable, however that is not always feasible.
My suggested fix would be to not evaluate getters for the logging purposes. They are likely to return new objects, perform a lot of work and thus slow down the execution, may change the state of the application, or may even throw. Only ever compare fields to avoid changing the state or executing code at all costs.
Using `Object.keys()` would improve the situation for properties defined on classes, but ideally "getter"s should not be called at all, no matter whether they are defined as own properties or prototype properties.
The current state requires complicated workarounds and causes hard to debug issues in complex applications. As it only affects dev-mode, I would love to see this improved in a bugfix so that people can upgrade to 19.2 | open | yGuy | 2025-11-13T09:39:02 | 2025-12-05T14:21:53 | null | [
"Type: Discussion"
] | null | 0 | {
"+1": 13,
"heart": 1
} | 12 | false | 2026-03-29T04:51:58.094982 | |||
facebook | react | 26,069 | I_kwDOAJy2Ks5dBFIK | false | Bug: Memory leak in react while focusing input elements | <!--
Please provide a clear and concise description of what the bug is. Include
screenshots if needed. Please test using the latest version of the relevant
React packages to make sure your issue has not already been fixed.
-->
I 've noticed that having an `input` element focused and unmounting its owner component doesn't actually free up this component from the js heap , I 've created this simple blnkr to demonstarate this which add a dummy `X` to a ref inside a component that has an input element
React version: 17.0.2
## Steps To Reproduce
1. Go to this [plnkr](https://plnkr.co/edit/IiqhT2JEvZcAJBMu?open=index.js&deferRun=1&preview)
2. Click Tab Two
3. Click the `Grow button` a few times
4. Take a heap snapshot
5. Click the Grow button a few more times
6. **IMPORTANT** Focus the input element by clicking into it
7. Click Tab One ( This unmount the component that own the ref to the list )
8. Take another heap snapshot ( after running the GC)
9. Compare between the two snapshots
10. The added string `X` will be found within the `concat check the screenshot
<img width="1507" alt="Screenshot 2023-01-28 at 1 13 12 AM" src="https://user-images.githubusercontent.com/28496859/215225175-f64cf1ee-f6b9-4dff-93d0-4fdebc3747e9.png">
<!--
Your bug will get fixed much faster if we can run your code and it doesn't
have dependencies other than React. Issues without reproduction steps or
code examples may be immediately closed as not actionable.
-->
Link to code example: can be found above
<!--
Please provide a CodeSandbox (https://codesandbox.io/s/new), a link to a
repository on GitHub, or provide a minimal code example that reproduces the
problem. You may provide a screenshot of the application if you think it is
relevant to your bug report. Here are some tips for providing a minimal
example: https://stackoverflow.com/help/mcve.
-->
## The current behavior
The current component that own the input element is still being there in the memory and doesn't free up its own resources afer being unmounted.
## The expected behavior
The current component shloud free up its resources whenever being unmounted
| open | reopened | Elwazer007 | 2023-01-27T23:38:32 | 2025-11-21T11:29:28 | null | [
"Type: Bug"
] | null | 0 | {} | 16 | false | 2026-03-29T04:53:12.461800 | ||
facebook | react | 26,052 | I_kwDOAJy2Ks5c13wc | false | [DevTools Bug]: event.metaKey + f to focus SearchInput doesn't work on Windows | ### Website or app
https://beta.reactjs.org/
### Repro steps
1. Open React Dev Tools -> Components
2. Try hitting (Windows Key) + f
On Windows 10/11, the Feedback Hub opens up. It will not focus on the search input.
I understand this probably works fine on Mac, but on Windows it'd be great to use a key that won't be intercepted by Windows. Like `Shift + f`. or `Ctrl + Alt + f`. Or, perhaps as soon as I start typing (unless I'm typing in another focused input). Or, if I press `/`. Something!
I thought at first DevTools didn't have a keyboard shortcut but then I looked at the source code and saw it uses [`metaKey`](https://developer.mozilla.org/en-US/docs/Web/API/KeyboardEvent/metaKey) which _totally_ doesn't work on Firefox in Windows.
### How often does this bug happen?
Every time
### DevTools package (automated)
_No response_
### DevTools version (automated)
_No response_
### Error message (automated)
_No response_
### Error call stack (automated)
_No response_
### Error component stack (automated)
_No response_
### GitHub query string (automated)
_No response_ | open | kamranayub | 2023-01-26T04:46:27 | 2025-11-07T05:46:22 | null | [
"Type: Bug",
"Status: Unconfirmed",
"Component: Developer Tools"
] | null | 0 | {} | 2 | false | 2026-03-29T04:51:53.754215 | |||
facebook | react | 35,070 | I_kwDOAJy2Ks7Wirzs | false | [Compiler Bug]: Update issue for nested objects | ### What kind of issue is this?
- [x] React Compiler core (the JS output is incorrect, or your app works incorrectly after optimization)
- [ ] babel-plugin-react-compiler (build issue installing or using the Babel plugin)
- [ ] eslint-plugin-react-hooks (build issue installing or using the eslint plugin)
- [ ] react-compiler-healthcheck (build issue installing or using the healthcheck script)
### Link to repro
https://playground.react.dev/#N4Igzg9grgTgxgUxALhHCA7MAXABAQzDATwF5dgAdDXXDfAWwWV0pABUEc2Aaa2-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-ou9EXBg3BBTAoQPgrFZRv8wNOwOGuKmG9nczOwAEfcHwFCIAB3ACSGBiMHoVjAKAsj4QaiAA
### Repro steps
If update button gets clicked, first street gets updated, but not the second one.
This seems to me unexpected behaviour.
### How often does this bug happen?
Every time
### What version of React are you using?
react 19
### What version of React Compiler are you using?
1.0 | closed | completed | JD2k | 2025-11-07T09:18:09 | 2025-11-07T16:36:21 | 2025-11-07T16:36:21 | [
"Type: Question"
] | null | 0 | {} | 2 | false | 2026-03-29T04:51:53.791643 | ||
facebook | react | 34,079 | I_kwDOAJy2Ks7DxEtu | false | Bug: renderToStaticMarkup fails with client-side code, suggesting a server-only context limitation | Description of the bug
When using renderToStaticMarkup to render a React component that contains client-side code (e.g., hooks like useState, event handlers like onClick, or browser-specific APIs), the function throws a runtime error. This behavior occurs even when the component is not actively being used in a browser environment.
To Reproduce
Steps to reproduce the behavior:
Create a simple React functional component that uses useState or useEffect without a dependency on a server-side framework.
Attempt to render this component using ReactDOMServer.renderToStaticMarkup().
The render process will fail with an error similar to "Hooks can only be called inside the body of a function component."
Expected behavior
The renderToStaticMarkup function should either:
a) Gracefully handle client-side code, perhaps by ignoring it or returning an empty string where it would be rendered, or
b) Provide a more explicit and descriptive error message that clearly states it cannot process client-side components.
Actual behavior
The function crashes with an error that is not immediately clear to the developer, suggesting a fundamental conflict between the static rendering process and the component's reliance on client-side state/effects.
Environment
React Version: [e.g., 18.2.0]
Node.js Version: [e.g., 18.12.1]
Browser: Not applicable, as the error occurs during server-side rendering.
Additional context
This limitation makes it difficult to use renderToStaticMarkup for tasks like generating static HTML for email templates or server-side caching if any of the rendered components contain even a small amount of client-specific logic. A clear explanation in the documentation or a more helpful error message would be beneficial. | closed | not_planned | Thanmaya-Dev | 2025-08-01T16:07:15 | 2025-11-07T08:07:42 | 2025-11-07T08:07:42 | [
"Status: Unconfirmed",
"Resolution: Stale"
] | null | 0 | {} | 3 | false | 2026-03-29T04:51:53.882846 | ||
facebook | react | 27,311 | I_kwDOAJy2Ks5vrsqA | false | Bug: Sudden scroll on react website | <!--
Sudden Scroll Behavior on React Website
-->
React version: 18.2.0
## Steps To Reproduce
1.Open the React website in a web browser.
2.Attempt to scroll up or down using the scroll wheel, trackpad, or scroll bar.
Link to code example:
<!--
https://github.com/facebook/react/assets/112266194/55229360-6c79-4dd0-9cae-c5aa0bad5641
-->
## The current behavior
When scrolling, the screen jumps suddenly to a different position on the page, causing an unpleasant and unexpected scrolling experience. This behavior makes it difficult to navigate and interact with the website content.
## The expected behavior
The scrolling should be smooth and fluid, allowing users to navigate through the website content without any sudden jumps.
| closed | completed | raghchawla131 | 2023-08-30T13:49:29 | 2025-11-09T10:48:23 | 2023-10-06T13:11:10 | [
"Status: Unconfirmed"
] | null | 0 | {
"+1": 3,
"eyes": 1
} | 5 | false | 2026-03-29T04:51:53.939351 | ||
facebook | react | 33,703 | I_kwDOAJy2Ks6-8UJ- | false | Bug: `react-hooks/exhaustive-deps` does more than checking deps and cannot be configured for intended behavior | I configured the `react-hooks/exhaustive-deps` eslint rule to lint my `useAsyncEffect` custom hook:
```
"react-hooks/exhaustive-deps": [
"warn",
{
additionalHooks: "(useAsyncEffect)",
},
],
```
but whenever I use it, it doesn't just make sure I include all dependencies in the deps array, it also errors with this:
```
Effect callbacks are synchronous to prevent race conditions. Put the async function inside:
useEffect(() => {
async function fetchData() {
// You can await here
const response = await MyAPI.getData(someId);
// ...
}
fetchData();
}, [someId]); // Or [] if effect doesn't need props or state
Learn more about data fetching with Hooks: https://react.dev/link/hooks-data-fetchingeslint[react-hooks/exhaustive-deps](https://github.com/facebook/react/issues/14920)
```
passing an async function to my custom useAsyncEffect hook is intended behavior.
the eslint rule `react-hooks/exhaustive-deps` is doing too much - these checks are separate and should be split into different rules, as per the single responsibility principle
at the least there should be a way to configure the rule so that I could use it with my custom hook | closed | not_planned | marcospgp | 2025-07-04T18:46:50 | 2025-11-08T15:04:41 | 2025-11-08T15:04:41 | [
"Status: Unconfirmed",
"Resolution: Stale"
] | null | 0 | {
"+1": 1
} | 3 | false | 2026-03-29T04:51:53.985061 | ||
facebook | react | 35,049 | I_kwDOAJy2Ks7WBUCU | false | [DevTools Bug] Cannot read properties of undefined (reading 'map') | ### Website or app
Profiler crashed
### Repro steps
Tried to record in the profiler.
Navigated to another screen in my react-native app, and profiler crash.
### How often does this bug happen?
Sometimes
### DevTools package (automated)
react-devtools-fusebox
### DevTools version (automated)
6.1.4-9fd4c09d68
### Error message (automated)
Cannot read properties of undefined (reading 'map')
### Error call stack (automated)
```text
at Sl (http://127.0.0.1:8081/debugger-frontend/third_party/react-devtools/react-devtools.js:41:891273)
at zo (http://127.0.0.1:8081/debugger-frontend/third_party/react-devtools/react-devtools.js:1:56181)
at xl (http://127.0.0.1:8081/debugger-frontend/third_party/react-devtools/react-devtools.js:1:78611)
at Xl (http://127.0.0.1:8081/debugger-frontend/third_party/react-devtools/react-devtools.js:1:89137)
at Wf (http://127.0.0.1:8081/debugger-frontend/third_party/react-devtools/react-devtools.js:1:140488)
at Vf (http://127.0.0.1:8081/debugger-frontend/third_party/react-devtools/react-devtools.js:1:140416)
at Bf (http://127.0.0.1:8081/debugger-frontend/third_party/react-devtools/react-devtools.js:1:140258)
at Mf (http://127.0.0.1:8081/debugger-frontend/third_party/react-devtools/react-devtools.js:1:136684)
at yd (http://127.0.0.1:8081/debugger-frontend/third_party/react-devtools/react-devtools.js:1:157182)
at MessagePort.C (http://127.0.0.1:8081/debugger-frontend/third_party/react-devtools/react-devtools.js:39:1432)
```
### Error component stack (automated)
```text
at Sl (http://127.0.0.1:8081/debugger-frontend/third_party/react-devtools/react-devtools.js:41:891063)
at div (<anonymous>)
at div (<anonymous>)
at bv (http://127.0.0.1:8081/debugger-frontend/third_party/react-devtools/react-devtools.js:41:966955)
at div (<anonymous>)
at div (<anonymous>)
at Dv (http://127.0.0.1:8081/debugger-frontend/third_party/react-devtools/react-devtools.js:41:968200)
at Pv (http://127.0.0.1:8081/debugger-frontend/third_party/react-devtools/react-devtools.js:41:972397)
at div (<anonymous>)
at lr (http://127.0.0.1:8081/debugger-frontend/third_party/react-devtools/react-devtools.js:41:775658)
at div (<anonymous>)
at Fv (http://127.0.0.1:8081/debugger-frontend/third_party/react-devtools/react-devtools.js:41:971851)
at div (<anonymous>)
at div (<anonymous>)
at div (<anonymous>)
at jo (http://127.0.0.1:8081/debugger-frontend/third_party/react-devtools/react-devtools.js:41:869288)
at http://127.0.0.1:8081/debugger-frontend/third_party/react-devtools/react-devtools.js:41:1089658
at el (http://127.0.0.1:8081/debugger-frontend/third_party/react-devtools/react-devtools.js:41:884838)
at http://127.0.0.1:8081/debugger-frontend/third_party/react-devtools/react-devtools.js:41:887488
at div (<anonymous>)
at div (<anonymous>)
at div (<anonymous>)
at tl (http://127.0.0.1:8081/debugger-frontend/third_party/react-devtools/react-devtools.js:41:887322)
at pc (http://127.0.0.1:8081/debugger-frontend/third_party/react-devtools/react-devtools.js:41:898959)
at Ud (http://127.0.0.1:8081/debugger-frontend/third_party/react-devtools/react-devtools.js:41:959076)
at Sd (http://127.0.0.1:8081/debugger-frontend/third_party/react-devtools/react-devtools.js:41:951587)
at jr (http://127.0.0.1:8081/debugger-frontend/third_party/react-devtools/react-devtools.js:41:789938)
at wt (http://127.0.0.1:8081/debugger-frontend/third_party/react-devtools/react-devtools.js:41:817279)
at Pf (http://127.0.0.1:8081/debugger-frontend/third_party/react-devtools/react-devtools.js:41:939000)
at oS (http://127.0.0.1:8081/debugger-frontend/third_party/react-devtools/react-devtools.js:41:1096424)
```
### GitHub query string (automated)
```text
https://api.github.com/search/issues?q=Cannot read properties of undefined (reading 'map') in:title is:issue is:open is:public label:"Component: Developer Tools" repo:facebook/react
``` | open | Amurmurmur | 2025-11-05T11:55:34 | 2025-11-07T15:52:57 | null | [
"Type: Bug",
"Status: Unconfirmed",
"Component: Developer Tools"
] | null | 0 | {} | 3 | false | 2026-03-29T04:51:54.047370 | |||
facebook | react | 34,011 | I_kwDOAJy2Ks7Cp_lI | false | Bug: Issue in the sidebar components with still using hsl | It looks like the sidebar component still use HSL in here:
```
const sidebarMenuButtonVariants = cva(
"peer/menu-button flex w-full items-center gap-2 overflow-hidden rounded-md p-2 text-left text-sm outline-hidden ring-sidebar-ring transition-[width,height,padding] hover:bg-sidebar-accent hover:text-sidebar-accent-foreground focus-visible:ring-2 active:bg-sidebar-accent active:text-sidebar-accent-foreground disabled:pointer-events-none disabled:opacity-50 group-has-data-[sidebar=menu-action]/menu-item:pr-8 aria-disabled:pointer-events-none aria-disabled:opacity-50 data-[active=true]:bg-sidebar-accent data-[active=true]:font-medium data-[active=true]:text-sidebar-accent-foreground data-[state=open]:hover:bg-sidebar-accent data-[state=open]:hover:text-sidebar-accent-foreground group-data-[collapsible=icon]:size-8! group-data-[collapsible=icon]:p-2! [&>span:last-child]:truncate [&>svg]:size-4 [&>svg]:shrink-0",
{
variants: {
variant: {
default: "hover:bg-sidebar-accent hover:text-sidebar-accent-foreground",
outline:
"bg-background shadow-[0_0_0_1px_hsl(var(--sidebar-border))] hover:bg-sidebar-accent hover:text-sidebar-accent-foreground hover:shadow-[0_0_0_1px_hsl(var(--sidebar-accent))]",
},
size: {
default: "h-8 text-sm",
sm: "h-7 text-xs",
lg: "h-12 text-sm group-data-[collapsible=icon]:p-0!",
},
},
defaultVariants: {
variant: "default",
size: "default",
},
},
);
```
The outline variant use hsl for wrapping the sidebar-border var, and for the hover effect.
Since tailwind 4 those vars are in oklch, so the `hsl()` need to be removed. | closed | not_planned | doneumark | 2025-07-26T15:12:32 | 2025-11-10T08:08:03 | 2025-11-10T08:08:03 | [
"Status: Unconfirmed",
"Resolution: Stale"
] | null | 0 | {
"+1": 1
} | 6 | false | 2026-03-29T04:51:54.087012 | ||
facebook | react | 35,073 | I_kwDOAJy2Ks7WsdmV | false | Bug: Errors thrown from "hidden" Activity boundaries shouldn't escape the boundary | <!--
Please provide a clear and concise description of what the bug is. Include
screenshots if needed. Please test using the latest version of the relevant
React packages to make sure your issue has not already been fixed.
-->
The provided repro is a contrived example, but a common example of where this is unexpected is if you were using Activity to render a previously visited page, and that page rendered a hook that throws an error if it's missing some sort of context.
In the case of Next.js, imagine we're grabbing route params via `useParams`, and we navigate to a route that no longer has that param available. When that subtree is switched to "hidden" mode, if any of the containing hooks/components threw an error when a param value is unavailable, that error would be triggered when navigating to the different page.
React version: 19.2.0
## Steps To Reproduce
1. Render an `<Activity>` boundary in "hidden" mode, and have it throw an error
2. Observe that even though the containing component is hidden, the error escapes and causes an error while rendering the "Visible" element.
<!--
Your bug will get fixed much faster if we can run your code and it doesn't
have dependencies other than React. Issues without reproduction steps or
code examples may be immediately closed as not actionable.
-->
Link to code example: https://codesandbox.io/p/devbox/distracted-poitras-yd3dkt
<!--
Please provide a CodeSandbox (https://codesandbox.io/s/new), a link to a
repository on GitHub, or provide a minimal code example that reproduces the
problem. You may provide a screenshot of the application if you think it is
relevant to your bug report. Here are some tips for providing a minimal
example: https://stackoverflow.com/help/mcve.
-->
## The current behavior
Throwing an error inside of a "hidden" Activity boundary triggers an error.
## The expected behavior
Throwing an error inside of a "hidden" Activity boundary should only trigger the error when the boundary is visible.
| closed | completed | ztanner | 2025-11-07T20:50:56 | 2025-11-07T23:23:18 | 2025-11-07T23:23:18 | [
"Status: Unconfirmed"
] | null | 0 | {} | 1 | false | 2026-03-29T04:51:54.153486 | ||
facebook | react | 34,111 | I_kwDOAJy2Ks7ENVVR | false | Bug: "Internal React error: Expected static flag was missing." | # "Internal React error: Expected static flag was missing. Please notify the React team."
This error message appeared after I created the component in the following code example. I pinpointed the issue down to the recursive nature of the component, which seems to confuse react in a way that it does not handle the early return *before* the calls to react hooks (which I placed there by accident) correctly.
```tsx
"use client";
import { useEffect, useState } from "react";
/* dummy type */
type ProductGroup = { id: string, name: string };
type SubGroupProps = {
depth: number,
label: string,
root: ProductGroup[],
action: (id: string) => Promise<ProductGroup[]>,
}
/* create recursive subgroups to select from */
function SubGroupFilter({ depth, label, root, action }: Readonly<SubGroupProps>) {
/* ------- BUG FROM HERE ------- */
/* early return */
if (!root.length)
return;
/* does not get catched!! */
const [index, setIndex] = useState<number>(0);
const [items, setItems] = useState<ProductGroup[]>([]);
useEffect(() => {
action(root[index].id).then(setItems);
}, [root, action, index, setItems]);
/* ---------- TO HERE ---------- */
return (
<>
<fieldset>
<legend>{label}</legend>
<select name="product-group" onChange={event => setIndex(event.currentTarget.selectedIndex)}>
{root.map(item => (
<option key={item.id} value={item.id}>
{item.name}
</option>
))}
</select>
</fieldset>
<SubGroupFilter depth={depth + 1} label={`Subgroup - ${root[index].name}`} root={items} action={action} />
</>
);
}
type SearchFormProps = {
root: ProductGroup[],
action: (id: string) => Promise<ProductGroup[]>,
}
export function SearchForm({ root, action }: Readonly<SearchFormProps>) {
return (
< >
<span >Search Form</span>
<form>
<SubGroupFilter depth={0} label="Product groups" root={root} action={action} />
<button className="button">search</button>
</form>
</>
);
}
```
```tsx
<SearchForm root={[{
id: "foo1",
name: "Foo1"
}, {
id: "foo2",
name: "Foo2"
}, {
id: "foo3",
name: "Foo3"
}, {
id: "foo4",
name: "Foo4"
}, {
id: "foo5",
name: "Foo5"
}]} action={async (id: string) => {
"use server";
if (id === "foo1" || id === "foo1-foo2" || id === "foo1-foo2-foo3")
return [{
id: `${id}-foo1`,
name: "Foo1"
}, {
id: `${id}-foo2`,
name: "Foo2"
}, {
id: `${id}-foo3`,
name: "Foo3"
}, {
id: `${id}-foo4`,
name: "Foo4"
}, {
id: `${id}-foo5`,
name: "Foo5"
}];
return [];
}}>
</SearchForm>
```
React version: 19.1.0
## Steps To Reproduce
1. create a new component with recursive subcomponents (like here with the filter subgroups)
2. make those recursive subcomponents call react hooks after an early return
In the example provided above the error message pops up every time a new subgroup appears, after selecting a parent group.
The issue does not appear when properly calling react hooks, i.e. move the return after the block of hook calls.
## The current and expected behavior
There is seemingly no change in behavior, nor any other issues except for the error message.
Hope this bug report somehow helps fixing the issue and make React a little better than it already is. | closed | not_planned | Scriptor25 | 2025-08-05T06:53:15 | 2025-11-10T08:08:02 | 2025-11-10T08:08:02 | [
"Status: Unconfirmed",
"Resolution: Stale"
] | null | 0 | {} | 2 | false | 2026-03-29T04:51:54.194832 | ||
facebook | react | 31,260 | I_kwDOAJy2Ks6aRTmw | false | Bug: Quickly selecting checkboxes on iOS doesn't work correctly | React version: 18.0.0 and 17.0.2
## Steps To Reproduce
1. Open https://4txdx2.csb.app/ (React 18) in a device running iOS in Safari / Chrome.
2. Quickly tap one checkbox followed by another one.
3. Notice that if you quickly select one checkbox followed by another one, it checks / unchecks the previous checkbox you were on on.
Link to code example:
React 18.0.0
[Sandbox](https://codesandbox.io/p/sandbox/react-18-checkbox-issue-4txdx2), [Preview](https://4txdx2.csb.app/)
React 17.0.2
[Sandbox](https://codesandbox.io/p/sandbox/react-17-checkbox-issue-yqgs6d), [Preview](https://yqgs6d.csb.app/)
React 16.4.0
[Sandbox](https://codesandbox.io/p/sandbox/react-16-checkbox-issue-jzsk96), [Preview](https://jzsk96.csb.app/)
## The current behavior
There appears to be some sort of race condition where tapping on a controlled / uncontrolled `<input type="checkbox">` quickly after tapping another checkbox updates the original checkbox you tapped rather than the checkbox that was just tapped. It appears to be a timing issue - if you wait long enough between taps, the events are fired on the correct elements.
This works in React 16 but not React 17 and 18.
## The expected behavior
The checkbox state should reflect the checked state of the checkbox you tapped, not the previous checkbox you tapped. | closed | not_planned | danielddb | 2024-10-15T09:46:53 | 2025-11-09T13:09:23 | 2025-11-09T13:09:23 | [
"Status: Unconfirmed",
"Resolution: Stale"
] | null | 0 | {
"+1": 5
} | 19 | false | 2026-03-29T04:51:54.256700 | ||
facebook | react | 32,576 | I_kwDOAJy2Ks6tmARR | false | Bug: Uncontrolled `<input />` element has value attribute removed | <!--
Please provide a clear and concise description of what the bug is. Include
screenshots if needed. Please test using the latest version of the relevant
React packages to make sure your issue has not already been fixed.
-->
React version: React 19
## Steps To Reproduce
1. Pass an `value={undefined}` as a prop to an `<input />` element
2. Observe the DOM and see the `value` attribute is absent
<!--
Your bug will get fixed much faster if we can run your code and it doesn't
have dependencies other than React. Issues without reproduction steps or
code examples may be immediately closed as not actionable.
-->
Link to code example:
[React 18](https://codesandbox.io/p/sandbox/34dw4q)
[React 19](https://codesandbox.io/p/sandbox/3p7hp9)
<!--
Please provide a CodeSandbox (https://codesandbox.io/s/new), a link to a
repository on GitHub, or provide a minimal code example that reproduces the
problem. You may provide a screenshot of the application if you think it is
relevant to your bug report. Here are some tips for providing a minimal
example: https://stackoverflow.com/help/mcve.
-->
## The current behavior
React 18 passed a _booleanish_ attribute i.e. `<input value />`. React 19 omitted the attribute altogether i.e. `<input />`. This is causing a lot of failures in our snapshot tests. I'm not sure if this behaviour is expected in React 19 and I can't find anything in the release notes apart from [this loosely related PR](https://github.com/facebook/react/pull/18513). If this behaviour is expected in 19, it would be great to mention that somewhere!
| closed | not_planned | imjordanxd | 2025-03-12T02:31:43 | 2025-11-10T08:08:07 | 2025-11-10T08:08:07 | [
"Status: Unconfirmed",
"Resolution: Stale"
] | null | 0 | {} | 8 | false | 2026-03-29T04:51:54.328487 | ||
facebook | react | 35,101 | I_kwDOAJy2Ks7XM1FV | false | [Compiler Bug]: Closure prop hoisted outside its lexical scope | ### What kind of issue is this?
- [x] React Compiler core (the JS output is incorrect, or your app works incorrectly after optimization)
- [ ] babel-plugin-react-compiler (build issue installing or using the Babel plugin)
- [ ] eslint-plugin-react-hooks (build issue installing or using the eslint plugin)
- [ ] react-compiler-healthcheck (build issue installing or using the healthcheck script)
### Link to repro
[Playground](https://playground.react.dev/#N4Igzg9grgTgxgUxALhAMygOzgFwJYSYAEaAFALICaA+gIIBKA4gKrkCiAcgCoCURwAHWJE4hMDiIAxCBCIBeIqT5yAfEQA8AIQCGMIgAcYEfXOBL5aqnSatOXAL5EA9CqFEiMBDljFpEIfYgADQgophoeADmKCB4ALb6EDASOACe+gj8RAAKADZQkXiYAPL6+GJEjmhGcUQA5ABG2g0IuQC0+vmFmG2e2rhtogl4uQgwTgAmeOJ1ANxCQqSCwk5OQ-oj2uWY5BATCMhEAiDaubnHAURgW9MRCGA5XUWl22A8s8HgABYQAO4AkpgcGNMKcwCg0GCEPYgA)
### Repro steps
In the following example, `() => MY_ARGUMENT` gets compiled to a top-level function `_temp`, but `MY_ARGUMENT` is unbound in `_temp`.
#### Input
```tsx
function f(MY_ARGUMENT) {
const Foo = () => <Bar prop={() => MY_ARGUMENT} />
return Foo
}
```
#### Output
```tsx
import { c as _c } from "react/compiler-runtime";
function f(MY_ARGUMENT) {
const Foo = () => …;
return Foo;
}
function _temp() {
return MY_ARGUMENT; // ❌ This is unbound!
}
```
This repros regardless of whether `Foo` is returned (though an unused `Foo` would probably be stripped out by a different optimizing pass). If `foo` is [lowercase](https://playground.react.dev/#N4Igzg9grgTgxgUxALhAMygOzgFwJYSYAEaAFALICaA+gIIBKA4gKrkCiAcgCoCURwAHWJE4hMDhIQIRALxFSfGQD4iAHgBCAQxhEADjAi6ZwBbJVU6TVpy4BfIgHolQokRgIcsYmilDbIABoQUUw0PABzFBA8AFtdCBgJHABPXQR+IgAFABsocLxMAHldfDEiezQDGKIAcgAjTTqEbIBaXVz8zBb3TVwW0Ti8bIQYBwATPHEagG4hIVJBYQcHAd0hzVLMcggxhGQiARBNbOzDvyIwDcmwhDAsjoLizbAeacDwAAsIAHcASUwcCNMMcwCg0CCELYgA), hoisting will not occur.
### How often does this bug happen?
Every time
### What version of React are you using?
"target": "19"
### What version of React Compiler are you using?
1.0.0 | open | sliminality | 2025-11-11T02:46:20 | 2025-11-11T02:46:42 | null | [
"Type: Bug",
"Status: Unconfirmed"
] | null | 0 | {} | 0 | false | 2026-03-29T04:51:54.378721 | |||
facebook | react | 35,104 | I_kwDOAJy2Ks7XPFwz | false | [Compiler Bug]: | ### What kind of issue is this?
- [ ] React Compiler core (the JS output is incorrect, or your app works incorrectly after optimization)
- [ ] babel-plugin-react-compiler (build issue installing or using the Babel plugin)
- [x] eslint-plugin-react-hooks (build issue installing or using the eslint plugin)
- [ ] react-compiler-healthcheck (build issue installing or using the healthcheck script)
### Link to repro
https://github.com/manNomi/rc-test
### Repro steps
## 🐞 Bug Report: `eslint-disable` incorrectly suppresses `incompatible-library` warning, causing silent memoization skip
### Summary
First and foremost, I want to express my sincere gratitude for the clarification provided in https://github.com/facebook/react/pull/35096. You were absolutely correct that the compiler automatically skips memoization for incompatible APIs, and I apologize for my earlier misunderstanding about needing `use no memo`.
I'm writing to humbly bring attention to what I believe may be an important edge case: an unrelated `eslint-disable` comment (such as for `exhaustive-deps`) appears to inadvertently suppress the `incompatible-library` warning. This transforms what should be a debuggable "automatic skip" (where developers are properly notified) into a silent skip scenario (where no notification occurs), which could potentially lead to confusion, especially for developers who are still learning React.
---
### The Two Scenarios
Please allow me to illustrate the concern through two scenarios:
#### Scenario 1: Expected Behavior (Debuggable)
1. A developer uses `useVirtualizer` (an incompatible API) within a custom hook
2. **Compiler:** Automatically and correctly skips memoization for this hook
3. **Linter:** Properly displays the `react-hooks/incompatible-library` warning
4. **Result:** The developer is informed that this hook is not memoized and can adjust their debugging approach accordingly
This behavior works wonderfully and as intended.
#### Scenario 2: Potential Issue (Silent Skip)
1. A developer uses `useVirtualizer` (an incompatible API) within a custom hook
2. In the same hook, they add a `useEffect` with an `eslint-disable-next-line react-hooks/exhaustive-deps` comment
3. **Compiler:** Skips memoization for this hook (due to the incompatible API)
4. **Linter:** The `eslint-disable-next-line react-hooks/exhaustive-deps` comment appears to also suppress the `react-hooks/incompatible-library` warning
5. **Result:** The hook is silently un-memoized without any warning, leaving the developer without a debugging signal
---
### Why This Matters
I understand this might be an edge case, and React's design philosophy encourages certain patterns over others. However, in real-world development, developers (especially those newer to React) sometimes write code in varied ways, which can lead to unexpected edge cases like this.
The particular concern is that **when code is extracted into a custom hook**, no error is thrown, yet only that custom hook loses its memoization—creating a subtle performance issue that's very difficult to trace.
The `incompatible-library` warning serves as a crucial debugging signal, informing developers that the compiler's automatic skip has occurred. When an unrelated `eslint-disable` rule inadvertently silences this critical signal, developers may believe their code is functioning correctly (since the linter appears clean), while the hook is actually un-memoized without their knowledge.
---
### Respectful Suggestion
If I may humbly suggest: perhaps an `eslint-disable` comment for one rule (such as `exhaustive-deps`) could be prevented from suppressing a different, critical warning (like `incompatible-library`)?
The `incompatible-library` warning might benefit from having higher precedence, displaying regardless of other `eslint-disable` comments, as it signals a fundamental state of the hook that developers should be made aware of.
---
### Closing Thoughts
I want to emphasize that I have tremendous respect and love for React—it has genuinely transformed my career and given me my livelihood. If the team believes this direction would be valuable, I would be more than happy to contribute and help implement a solution to the best of my abilities.
Thank you so much for taking the time to consider this report, and for all the incredible work you do on React. The React Compiler is an amazing advancement for the ecosystem, and I'm excited to see it evolve.
With sincere appreciation and respect.
### How often does this bug happen?
Every time
### What version of React are you using?
19.2.0
### What version of React Compiler are you using?
1.0.0 | closed | completed | nomiupsight | 2025-11-11T07:28:10 | 2025-11-11T07:28:27 | 2025-11-11T07:28:27 | [
"Type: Bug",
"Status: Unconfirmed"
] | null | 0 | {} | 0 | false | 2026-03-29T04:51:54.437833 | ||
facebook | react | 35,000 | I_kwDOAJy2Ks7UQx_c | false | Activity mode=“hidden” does not hide nested portals | React's `<Activity>` component does not hide portals when they are nested under another element. Only direct portal children are hidden, causing inconsistent behavior.
React version: 19.2.0
## Steps To Reproduce
1. Wrap a parent `<div>` inside `<Activity mode="hidden">`
2. Inside that `<div>`, render any regular child and a portal via `createPortal(...)`
3. Observe: regular child is hidden but the portal remains visible
4. Remove the additional wrapper and observe that the portal is now hidden correctly
Link to code example: https://codesandbox.io/p/sandbox/hwt3pl
Minimal reproduction:
```jsx
import { createPortal } from "react-dom";
import { Activity } from "react";
export default function App() {
return (
<Activity mode="hidden">
<div>
<h1>Hello</h1>
{createPortal(<p>Portal child (should be hidden)</p>, document.body)}
</div>
</Activity>
);
}
```
## The current behavior
Portals nested within another element inside remain visible when Activity is in “hidden” mode.
## The expected behavior
All children of an element, including portals nested at any depth, should be hidden and restored consistently when using “hidden” mode.
---
## Notes
Not sure if this is a bug or expected behavior, though. But I feel like we should either find a way to hide the portal or update the docs to mention this on the "unwanted side effects" section. Thoughts? | closed | completed | ceolinwill | 2025-10-28T11:41:56 | 2025-11-10T19:39:41 | 2025-11-10T19:39:41 | [
"Status: Unconfirmed"
] | null | 0 | {
"+1": 11,
"heart": 2
} | 3 | false | 2026-03-29T04:51:54.459422 | ||
facebook | react | 33,555 | I_kwDOAJy2Ks68Atv6 | false | [DevTools Bug]: Unnecessary Re-render Highlight on Sibling Component in React DevTools | ### Website or app
https://codesandbox.io/p/sandbox/agitated-monad-h4ykn2?file=%2Fsrc%2FApp.js%3A12%2C13
### Repro steps
* Open the provided CodeSandbox
* Click the "Open in new Tab" button in the sandbox preview.
* Open the React DevTools panel.
* In the DevTools settings, enable "Highlight updates when components render".
* Observe that the Test button(B) is highlighted every time the timer value changes.
### How often does this bug happen?
Every time
### DevTools package (automated)
_No response_
### DevTools version (automated)
_No response_
### Error message (automated)
_No response_
### Error call stack (automated)
```text
```
### Error component stack (automated)
```text
```
### GitHub query string (automated)
```text
``` | open | yuki2006 | 2025-06-17T17:53:36 | 2025-11-12T14:30:38 | null | [
"Type: Bug",
"Status: Unconfirmed",
"Component: Developer Tools"
] | null | 0 | {
"eyes": 1,
"heart": 2
} | 3 | false | 2026-03-29T04:51:54.499598 | |||
facebook | react | 32,743 | I_kwDOAJy2Ks6vqyeZ | false | Bug: React Dev Tools does not have a Google Chrome Store Established Publisher badge | The author for the React Dev Tools on Chrome Store is not an Established Publisher. Due to this, my security team will not approve our use for this extension. They point to other extensions that meet this qualification and I was shocked to see that the chrome extension's author is not an Established Publisher.
React version:
## Steps To Reproduce
1. visit this link: https://chromewebstore.google.com/detail/react-developer-tools/fmkadmapgofadopljbjfkapdkoienihi
2. Notice the missing Established Publisher Badge found to the left of the extension name, like in this example: https://chromewebstore.google.com/detail/keeper%C2%AE-password-manager/bfogiafebfohielmmehodmfbbebbbpei
Link to code example:
<!--
Please provide a CodeSandbox (https://codesandbox.io/s/new), a link to a
repository on GitHub, or provide a minimal code example that reproduces the
problem. You may provide a screenshot of the application if you think it is
relevant to your bug report. Here are some tips for providing a minimal
example: https://stackoverflow.com/help/mcve.
-->
## The current behavior
The badge does not appear to the left of the extension name.
## The expected behavior
The badge should appear to the left of the extension name.
| closed | not_planned | sygint | 2025-03-25T17:09:30 | 2025-11-11T19:04:41 | 2025-11-11T19:04:41 | [
"Status: Unconfirmed",
"Resolution: Stale"
] | [
"hoxyq"
] | 0 | {} | 6 | false | 2026-03-29T04:51:54.546824 | ||
facebook | react | 35,105 | I_kwDOAJy2Ks7XPGgl | false | [Compiler Bug]: eslint-disable incorrectly suppresses incompatible-library warning, causing silent memoization skip | ### What kind of issue is this?
- [ ] React Compiler core (the JS output is incorrect, or your app works incorrectly after optimization)
- [ ] babel-plugin-react-compiler (build issue installing or using the Babel plugin)
- [x] eslint-plugin-react-hooks (build issue installing or using the eslint plugin)
- [ ] react-compiler-healthcheck (build issue installing or using the healthcheck script)
### Link to repro
https://github.com/manNomi/rc-test
### Repro steps
## 🐞 Bug Report: `eslint-disable` incorrectly suppresses `incompatible-library` warning, causing silent memoization skip
### Summary
First and foremost, I want to express my sincere gratitude for the clarification provided in https://github.com/facebook/react/pull/35096. You were absolutely correct that the compiler automatically skips memoization for incompatible APIs, and I apologize for my earlier misunderstanding about needing `use no memo`.
I'm writing to humbly bring attention to what I believe may be an important edge case: an unrelated `eslint-disable` comment (such as for `exhaustive-deps`) appears to inadvertently suppress the `incompatible-library` warning. This transforms what should be a debuggable "automatic skip" (where developers are properly notified) into a silent skip scenario (where no notification occurs), which could potentially lead to confusion, especially for developers who are still learning React.
---
### The Two Scenarios
Please allow me to illustrate the concern through two scenarios:
#### Scenario 1: Expected Behavior (Debuggable)
1. A developer uses `useVirtualizer` (an incompatible API) within a custom hook
2. **Compiler:** Automatically and correctly skips memoization for this hook
3. **Linter:** Properly displays the `react-hooks/incompatible-library` warning
4. **Result:** The developer is informed that this hook is not memoized and can adjust their debugging approach accordingly
This behavior works wonderfully and as intended.
#### Scenario 2: Potential Issue (Silent Skip)
1. A developer uses `useVirtualizer` (an incompatible API) within a custom hook
2. In the same hook, they add a `useEffect` with an `eslint-disable-next-line react-hooks/exhaustive-deps` comment
3. **Compiler:** Skips memoization for this hook (due to the incompatible API)
4. **Linter:** The `eslint-disable-next-line react-hooks/exhaustive-deps` comment appears to also suppress the `react-hooks/incompatible-library` warning
5. **Result:** The hook is silently un-memoized without any warning, leaving the developer without a debugging signal
---
### Why This Matters
I understand this might be an edge case, and React's design philosophy encourages certain patterns over others. However, in real-world development, developers (especially those newer to React) sometimes write code in varied ways, which can lead to unexpected edge cases like this.
The particular concern is that **when code is extracted into a custom hook**, no error is thrown, yet only that custom hook loses its memoization—creating a subtle performance issue that's very difficult to trace.
The `incompatible-library` warning serves as a crucial debugging signal, informing developers that the compiler's automatic skip has occurred. When an unrelated `eslint-disable` rule inadvertently silences this critical signal, developers may believe their code is functioning correctly (since the linter appears clean), while the hook is actually un-memoized without their knowledge.
---
### Respectful Suggestion
If I may humbly suggest: perhaps an `eslint-disable` comment for one rule (such as `exhaustive-deps`) could be prevented from suppressing a different, critical warning (like `incompatible-library`)?
The `incompatible-library` warning might benefit from having higher precedence, displaying regardless of other `eslint-disable` comments, as it signals a fundamental state of the hook that developers should be made aware of.
---
### Closing Thoughts
I want to emphasize that I have tremendous respect and love for React—it has genuinely transformed my career and given me my livelihood. If the team believes this direction would be valuable, I would be more than happy to contribute and help implement a solution to the best of my abilities.
Thank you so much for taking the time to consider this report, and for all the incredible work you do on React. The React Compiler is an amazing advancement for the ecosystem, and I'm excited to see it evolve.
With sincere appreciation and respect.
### How often does this bug happen?
Every time
### What version of React are you using?
19.2.0
### What version of React Compiler are you using?
1.0.0 | open | manNomi | 2025-11-11T07:29:12 | 2025-11-13T00:21:03 | null | [
"Type: Bug",
"Status: Unconfirmed"
] | null | 0 | {} | 3 | false | 2026-03-29T04:51:54.587163 | |||
facebook | react | 35,055 | I_kwDOAJy2Ks7WRGF8 | false | [DevTools Bug]: DevTools in Edge constantly display "Looks like this page doesn't have React, or it hasn't been loaded yet." | ### Website or app
https://github.com/mgrosperrin/repro_issues/tree/react-dev-tools-edge
### Repro steps
1. Run `npm run dev`
2. In Edge, open the DevTools and go to the "Components" or "Profiler" tabs
Additionaly when I go back to the "Sources" built-in tabs, it was completly empty like if he crashed.
### How often does this bug happen?
Every time
### DevTools package (automated)
_No response_
### DevTools version (automated)
_No response_
### Error message (automated)
_No response_
### Error call stack (automated)
```text
```
### Error component stack (automated)
```text
```
### GitHub query string (automated)
```text
``` | closed | completed | mgrosperrin | 2025-11-06T09:02:02 | 2025-11-13T05:27:59 | 2025-11-09T14:04:07 | [
"Type: Bug",
"Status: Unconfirmed",
"Component: Developer Tools"
] | null | 0 | {
"+1": 4
} | 14 | false | 2026-03-29T04:51:54.621559 | ||
facebook | react | 31,285 | I_kwDOAJy2Ks6ax-IT | false | [DevTools Bug]: Phantom re-renders on sibling <label> components | ### Website or app
https://stackblitz.com/edit/react-devtools-bug?file=src%2FApp.jsx
### Repro steps
1. Start profiler
2. Input into "Component with state"
3. Both of "ComponentWithState" and "AnotherReactComponent" were re-rendered will be shown in the report but why?
### How often does this bug happen?
Every time
### DevTools package (automated)
_No response_
### DevTools version (automated)
_No response_
### Error message (automated)
_No response_
### Error call stack (automated)
_No response_
### Error component stack (automated)
_No response_
### GitHub query string (automated)
_No response_ | open | jindong-zhannng | 2024-10-18T08:08:40 | 2025-11-13T11:33:32 | null | [
"Type: Bug",
"Status: Unconfirmed",
"Component: Developer Tools"
] | null | 0 | {
"+1": 2
} | 10 | false | 2026-03-29T04:51:54.667534 | |||
facebook | react | 35,118 | I_kwDOAJy2Ks7XhXe9 | false | [Compiler Bug]: setter not correctly recognized as setter, required in `useCallback` dependency array | ### What kind of issue is this?
- [x] React Compiler core (the JS output is incorrect, or your app works incorrectly after optimization)
- [ ] babel-plugin-react-compiler (build issue installing or using the Babel plugin)
- [ ] eslint-plugin-react-hooks (build issue installing or using the eslint plugin)
- [ ] react-compiler-healthcheck (build issue installing or using the healthcheck script)
### Link to repro
https://playground.react.dev/#N4Igzg9grgTgxgUxALhAHQHYEsC2AHCGAFwAIAqEgQzBICUFK5SAzGCHEtEGBprgbkwIAHgWIlmUDEywQMJKGAQBlKACMwcGFjxFZGABSYSJCLv00AvCWABfTAEobmYyThywpYK5NgA1joANK62JNZmeh6CGK7uGJ4kPFoMRAhhJAZOlgB8Gf46JAD8ChgAJgjMWBgIpSTINrYOLvIkADYIpADaEBoIMABulGrtgSRKRADyvQND7QC66fSMRAB0iipElKlGLSYRFiv5eEUkGFCtrXWJCMlbCK4O0bEepEk8d-TMi7yr658GbxSCEernWAElIDhCHgABZYOAAGUoAE9oEQAKLMZg3IgGTJhXLeXbXW6pT4rOCwHgYUjWQF3aImRpPFpYL4GACERxIADIeSQOT0lDNhsDnMTxlNhYNRXihX0Ze10vTtg41YySPYYi04gkeJ5KOJrEsmGslABhSgXNSMPx4rKEnwkNkZeUi9pOIkmb1jDpShWzBAAm7vMkVClUhA0zIg4lapmjbrTRUIOaxkwAegzJE6KzzCwAKjC0lVsTAeLVyngo+VpMiSAB3agkAAGkuTgZbozUUFIRGLY2g8DSVZrUbgWAQNAbfTSSelgbmKxIYIwZYrJFKbLLUdIo7K4-r-co8kgsEQK2aJh4RFg8hNvyUAFkEFD7QSMsBrgbxI1E-rNmINNoi1EBAhAOJKgAcxQEBcDEPtkWrGwSAABVaKAoKqCZzA8TUJDYDgAHIbTUBBWgAWjwDCsIwCj3iYCj3HwLB2hgDMt08IiWQML0SCzZi8FYrZ9CfCBynqLgrVaLhMFCMARLASopzQmjsNw+JHjA8AYQgBtV1SGAMCtMAUCIGAoAQWwgA
### Repro steps
In the example,
```ts
const restart = React.useCallback(() => {
if (observable) {
setObservable(recreateRef.current());
}
}, [observable]);
```
should be fine, but the compiler complains that `setObservable` is missing from the dependency array, even though it is a`React.useState` return value from the same hook.
> The inferred dependency was `setObservable`, but the source dependencies were [observable]. Inferred different dependency than source.
Apologies for the weird reproduction - I couldn't reproduce this from scratch, so it's a minimized variant of where we're seeing this in the wild: https://github.com/apollographql/apollo-client/blob/a43583377ccaae60b58bc45edfc88f32bed74a46/src/react/hooks/useSubscription.ts#L375
I know that on it's own, the reproduction doesn't make a lot of sense, but if I leave any of the leftover building blocks out, everything works nicely.
### How often does this bug happen?
Every time
### What version of React are you using?
19.1.0, but likely irrelevant - this is a library with peerDependency 17-19
### What version of React Compiler are you using?
1.0.0/playground | open | phryneas | 2025-11-12T10:07:25 | 2025-11-13T13:24:23 | null | [
"Type: Bug",
"Status: Unconfirmed"
] | null | 0 | {
"+1": 1,
"heart": 7
} | 3 | false | 2026-03-29T04:51:54.727618 | |||
facebook | react | 35,142 | I_kwDOAJy2Ks7YMVzF | false | Bug: recovery my settlements | <!--
Please provide a clear and concise description of what the bug is. Include
screenshots if needed. Please test using the latest version of the relevant
React packages to make sure your issue has not already been fixed.
-->
React version:
## Steps To Reproduce
1.
2.
<!--
Your bug will get fixed much faster if we can run your code and it doesn't
have dependencies other than React. Issues without reproduction steps or
code examples may be immediately closed as not actionable.
-->
Link to code example:
<!--
Please provide a CodeSandbox (https://codesandbox.io/s/new), a link to a
repository on GitHub, or provide a minimal code example that reproduces the
problem. You may provide a screenshot of the application if you think it is
relevant to your bug report. Here are some tips for providing a minimal
example: https://stackoverflow.com/help/mcve.
-->
## The current behavior
## The expected behavior
| closed | not_planned | impeccableinfinity | 2025-11-14T20:46:40 | 2025-11-14T22:45:34 | 2025-11-14T22:45:34 | [
"Status: Unconfirmed"
] | null | 0 | {} | 0 | false | 2026-03-29T04:51:54.759250 | ||
facebook | react | 35,116 | I_kwDOAJy2Ks7XfhM7 | false | 💡 Platform Suggestion: Bounty System for React Library | Hi React Library team! 👋
I wanted to share **Roxonn** - a decentralized bounty platform that could help accelerate development.
**What is Roxonn?**
✅ Turn GitHub issues into crypto bounties (XDC, USDC, ROXN)
✅ Notify 300+ developers instantly
✅ Auto-pay when PRs merge via blockchain
✅ Zero crypto setup needed
**Quick flow:**
1. Register repo (GitHub App install)
2. Fund pool with USDC (stable pricing)
3. Assign bounties to features/bugs
4. PR merged → automatic payment
**Why teams use it:**
- Access to active developer network
- **Only 1% total platform fee**
- Transparent blockchain payments
Learn more: **https://roxonn.com**
*No pressure - just sharing a resource!* | closed | not_planned | dineshroxonn | 2025-11-12T07:40:45 | 2025-11-16T08:49:50 | 2025-11-16T08:49:50 | null | null | 0 | {
"heart": 2
} | 0 | false | 2026-03-29T04:51:54.854375 | ||
facebook | react | 35,155 | I_kwDOAJy2Ks7YZUn9 | false | Test | Test issue | closed | completed | MilosMosovsky | 2025-11-16T12:59:21 | 2025-11-16T13:00:31 | 2025-11-16T13:00:31 | null | null | 0 | {} | 0 | false | 2026-03-29T04:51:54.878862 | ||
facebook | react | 35,156 | I_kwDOAJy2Ks7YZVLm | false | Test | Test issue | closed | completed | MilosMosovsky | 2025-11-16T13:01:32 | 2025-11-16T13:01:59 | 2025-11-16T13:01:59 | null | null | 0 | {} | 0 | false | 2026-03-29T04:51:54.901826 | ||
facebook | react | 35,149 | I_kwDOAJy2Ks7YQYLD | false | Bug: React is good...! | REACT is gooooodddd!!!!! | closed | completed | SabziDev | 2025-11-15T06:23:53 | 2025-11-15T06:28:27 | 2025-11-15T06:28:27 | [
"Status: Unconfirmed"
] | null | 0 | {} | 0 | false | 2026-03-29T04:51:54.943486 | ||
facebook | react | 26,906 | I_kwDOAJy2Ks5n5XoK | false | Bug: there is no renderToReadableStream function in react-dom@18.2.0 in Node.js | I try to use `renderToReadableStream` function from `react-dom/server` in Node.js but there is no such function in package.
React version: 18.2.0
Node.js version 16.15.0 or 18.16.0
## Steps To Reproduce
1. create `index.mjs` file
2. insert code:
```js
import { renderToString, renderToReadableStream } from "react-dom/server";
console.log({ renderToString, renderToReadableStream });
```
3. run `node index.mjs` in terminal
Link to code example:
[StackBlitz](https://stackblitz.com/edit/node-yn12fy?file=index.mjs)
## The current behavior
Error in terminal:
```sh
file:///Users/user/Projects/node-fetch-api-server/index.mjs:1
import { renderToString, renderToReadableStream } from "react-dom/server";
^^^^^^^^^^^^^^^^^^^^^^
SyntaxError: Named export 'renderToReadableStream' not found. The requested module 'react-dom/server' is a CommonJS module, which may not support all module.exports as named exports.
CommonJS modules can always be imported via the default export, for example using:
import pkg from 'react-dom/server';
const { renderToString, renderToReadableStream } = pkg;
at ModuleJob._instantiate (node:internal/modules/esm/module_job:124:21)
at async ModuleJob.run (node:internal/modules/esm/module_job:190:5)
Node.js v18.16.0
```
## The expected behavior
Node.js suupports web streams in experimental mode.
The package `react-dom/server` must have a function `renderToReadableStream` according to the documentation:
https://react.dev/reference/react-dom/server/renderToReadableStream | closed | completed | krutoo | 2023-06-06T05:08:00 | 2025-11-17T08:55:06 | 2025-11-17T08:55:06 | [
"Status: Unconfirmed"
] | null | 0 | {
"+1": 18
} | 29 | false | 2026-03-29T04:51:54.968542 | ||
facebook | react | 33,153 | I_kwDOAJy2Ks61t5M_ | false | Bug: React error #421 isn't reported anymore | React error [#421](https://react.dev/errors/421) isn't reported anymore (Suspense boundary received an update before it finished hydrating), which makes debugging this error quite challenging.
When Suspense receives an update while hydration, React switches to client rendering, which causes Suspense to show the fallback for a brief moment of time, creating a blink that deteriorates the UX.
React version: **19.1.0**
## Steps To Reproduce
1. clone the repository
2. run `npm i`
3. run `npm run dev`
4. open http://localhost:5173/
5. make sure you see the second div do a blink during hydration
Link to code example:
https://github.com/stanislav-halyn/react-19-missing-hydration-errors
## The current behavior
No hydration error reported neither to the console or `onRecoverableError`.
## The expected behavior
Hydration error should be reported either to the console or `onRecoverableError`. | closed | not_planned | stanislav-halyn | 2025-05-08T11:30:55 | 2025-11-17T10:06:43 | 2025-11-17T10:06:43 | [
"Status: Unconfirmed",
"Resolution: Stale"
] | null | 0 | {
"+1": 1
} | 4 | false | 2026-03-29T04:51:55.037962 | ||
facebook | react | 31,722 | I_kwDOAJy2Ks6iw0Ag | false | Bug: `react-hooks/rules-of-hooks` does not allow `_Component` names underscore prefix for 'private naming convention' | <!--
Please provide a clear and concise description of what the bug is. Include
screenshots if needed. Please test using the latest version of the relevant
React packages to make sure your issue has not already been fixed.
-->
Latest version `eslint-plugin-react-hooks@5.1.0`
## Steps To Reproduce
```
export function _SomeComponentWeNotWantUsedOutsideThisDirectory() {
const result = useSomeHook();
}
```
## The current behavior
Produces eslint error
> React Hook "useSomeHook" is called in function "_SomeComponentWeNotWantUsedOutsideThisDirectory" that is neither a React function component nor a custom React Hook function. React component names must start with an uppercase letter. React Hook names must start with the word "use". eslint([react-hooks/rules-of-hooks](https://reactjs.org/docs/hooks-rules.html))
## The expected behavior
No eslint warning, because we have clearly indicated it is a component using the uppercase letter after the underscore prefix.
## Proposed solution
https://github.com/facebook/react/blob/4a8fc0f92e0f75257962522b51a938bf4dfda77a/packages/eslint-plugin-react-hooks/src/RulesOfHooks.js#L49
could be easily changed to
```ts
return node.type === 'Identifier' && /^_?[A-Z]/.test(node.name);
```
I have thought about whether it should be a config option but I'm leaning towards 'less is more'. Either way, I'm willing to make PR for this and test coverage but need confirmation you guys will take this into consideration before I put in the work. | open | 700software | 2024-12-10T17:12:34 | 2025-11-13T17:28:36 | null | [
"Type: Discussion",
"Component: ESLint Rules"
] | null | 0 | {
"+1": 6,
"eyes": 1
} | 6 | false | 2026-03-29T04:51:55.079094 | |||
facebook | react | 35,040 | I_kwDOAJy2Ks7V1kXV | false | Bug: Getting `Cannot access refs during render` error from event handler. | So on react official documentations, I found, it recommend to avoid direct read or write during render on ref.
Here is the docs link (under pitfall)-
https://react.dev/reference/react/useRef
And here I also found that ref is safe to use on event handler or effect.
`You can read or write refs from event handlers or effects instead.`
So, we can read `ref.current` from a onSubmit handler as it is a event handler. But when I combine it with `react-hook-form` I am getting the lint error with following context-
```
error Error: Cannot access refs during render
React refs are values that are not needed for rendering. Refs should only be accessed outside of re
nder, such as in event handlers or effects. Accessing a ref value (the `current` property) during r
ender can cause your component not to update as expected (https://react.dev/reference/react/useRef)
```
So What I am doing on my code(see code)-
```javascript
const onSubmit: SubmitHandler<ReceiptFormInput> = async (value) => {
let sigUrl: string;
if (sig) {
const { blob } = await upload(ref.current?.toFile());
sigUrl = blob?.url || ""
} else {
sigUrl = item.signature
}
if (template) {
mutate({
...value,
templateId: template.templateId,
previewUrl: template.previewUrl,
signature: sigUrl,
id: item.id
});
}
}
```
this onSubmit handler is triggering from `form`
```javascript
<form onSubmit={handleSubmit(onSubmit)}>
...
</form>
```
So here handleSubmit is coming from `react-hook-form` to disable default behavior.
We already discussed about this issues from `nextjs` discussion. Following is the link of that discussion
https://github.com/vercel/next.js/discussions/85677 | closed | completed | siamahnaf | 2025-11-04T17:36:26 | 2025-11-14T17:00:35 | 2025-11-14T17:00:35 | [
"Status: Unconfirmed"
] | null | 0 | {} | 5 | false | 2026-03-29T04:51:55.143574 | ||
facebook | react | 34,194 | I_kwDOAJy2Ks7Fv9SC | false | [Compiler Bug]: Incorrect behavior when using potentially undefined value in function without optional chaining | ### What kind of issue is this?
- [x] React Compiler core (the JS output is incorrect, or your app works incorrectly after optimization)
- [ ] babel-plugin-react-compiler (build issue installing or using the Babel plugin)
- [ ] eslint-plugin-react-compiler (build issue installing or using the eslint plugin)
- [ ] react-compiler-healthcheck (build issue installing or using the healthcheck script)
### Link to repro
https://playground.react.dev/#N4Igzg9grgTgxgUxALhASwLYAcIwC4AEwBAilAjAJ4DCANmggHZ4A0BUYCZFlBAvgQBmMCBgIByAAJ4AhozCy4AawD0MBDLh4AtAEdyVcQG4AOo0w58BAIIAPNBDBtidh2ABKCfQgXUIjQTQAc34hETFxGXtHYzMLXEJPTVYidk5qGVpaACNNJVDhUQl1ZNjGMzh-BTSEag48UQBJeVlGRAIAXgIACgBKToA+IjMCAnU8WEYeyoDg5Btojy9yX39AoLYILDwHeQB+eddHT29V2aDe+YAFcLROAB45SiGOoeAR0YIZ6shYdq6jmAAHQZNoIWgAFQgSiYQN+8AQfVM5Smn2+hCwt04nQWbm671RnwIQJJM3WLA+RJJQK2OyqFMJaLkiEh0KY83hiCBDRhjAZRNGAAsNAATChgeYEgWfalk4JA4UyMUwJyUgXU2m7MB7BWi8X8gV8A2jPi9bnCxjdfEEEUyWT8fqvG12mS9ZFqlQqAjSMDaYKMXAINWY0R3BBAuDM8E4vqDYaM0ac8ORsG0bribhUAgAdxkYC+UayCBFBGyvCSWlIBko4jdar47sZ40mBBDGDDyJNyIbZjMKgAVP2CABxJgUO3Fr4QMUEADKEOs7ghBH7KgqVUIHAQo7wABEEAA3NCIXcugASEGhMcdbw+6IIQQQe8Px4Qp-tXS3dQUTRaUaRvaos2MBTF+mQ5Hk3Rqt0OBgGgdLmCKHJ4DAaCMBsBBMDseCUMhqHoZs2xagcBBPGwcFBIwmQkdY2QJLO-qZDe8bSsBUyPs+R4ni6UEJp8xCwLQ8wAAYqIJKgACTALB8G7GgIp8JJwBYfBlB8MJbAYE+grTvM4jDgAohC4jkYxtD8MaAqalUaqjHWjJGmqADaHH7lxb4ugAuh89k9ii94cTubmvu+MiZpQADSCC8F0MGOLJ-jyXhaEYSpOHJehzFSmMT4tk5oniVJMkIfJilSWlanCZ53aNveW5BS+3GyOFADyRFVDGHzFXJSEEAo+EbB8FUZYNqLWfskoEN4VAkU8Rg5aceCzYwvB8GYWV3huqTTbhU3Vm1CFOAtKx4PM6iLQdWqhF0435nsexEH5m0tHtPBRTFao7ZdVQ6jt70EPdD5Pg17mheF71xXBJUimwFX2c91SuY1HkfjUIMheel5KABKKjPeO0AGKMPMTwxsQFFUeZfAkytN5qkjoM8d1iUw5hzCqWw50nd98imZRTGNqMbEsZ8f3RZZhN8mqTAyNktDFvMACEiuQwliEEAAZBrbPYZQvSWdSX3tbzHx+U9jAILYliEIIUBtAhaNPsFTUyLxozM4hI0GsNfUoSlBq3SR2WjDty2UJ2nxcz4S00+HpsGjtdAMMwJHhUnWHrbHIv4-txv5p+nDoy7rV56r0Ow+zOGEYd8OojnPA4lu4XdEbh1sIn9BYbXQu5SBr1UN2faDiOY4wBOBAGQAcruK5rv5W0ALI0KIOAW8w8wVngQIE9Q15xtl97ELasgctACLXY7nGvumkBaY0IomRISa1hH94YJQBN21oux706wcEGgQQPQkyKzhKIBALU8DChgDcLYFAcL9H-njKoEB5ZAloBAII6ZZYQAPAgAB+YoH4JDHLBAGAX71lNhHYW9wRRoAPAQfwSdlAdGAO-T+9tdh8AGMAJMOpLZYDQFQd8+CHpJiBAIoRlAREED0tocQfB7gqDoQeAYNVAICISDaBAggZBQFoIQJefhsD+CwsiEAfAgA
### Repro steps
When you run following code it will cause following error despite me not calling function that is calling this part of the code.
`Cannot read properties of undefined (reading 'someOtherProperty')`
When compiler is not enabled for this code this component will work fine on initial render. I used onClick event handler in the example but you can use any event handler and this error will happen. I included my generated API because it uses react-query and I'm not 100% sure if problem might be in it since I tried hard-coding some data and problem didn't happen in that case so problem could be in it too.
My use case is that I have checked the possible undefined value somewhere before calling the function and have determined that calling the function is safe and so use TS non-null assertion.
### How often does this bug happen?
Every time
### What version of React are you using?
19.1.0
### What version of React Compiler are you using?
19.1.0-rc.2 | open | SampsaKaskela | 2025-08-13T09:40:47 | 2025-11-17T19:21:30 | null | [
"Type: Bug",
"Component: React Compiler"
] | null | 0 | {
"+1": 8
} | 2 | false | 2026-03-29T04:51:55.189261 | |||
facebook | react | 35,157 | I_kwDOAJy2Ks7YfPYg | false | [Compiler]: Detect mutation of state values using type information | ### What kind of issue is this?
- [ ] React Compiler core (the JS output is incorrect, or your app works incorrectly after optimization)
- [ ] babel-plugin-react-compiler (build issue installing or using the Babel plugin)
- [x] eslint-plugin-react-hooks (build issue installing or using the eslint plugin)
- [ ] react-compiler-healthcheck (build issue installing or using the healthcheck script)
### Link to repro
https://playground.react.dev/#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
### Repro steps
Ref. https://react.dev/reference/eslint-plugin-react-hooks/lints/immutability
In the above article, some invalid patterns are introduced, however, the playground do not throw any errors for the violations.
Is this the limitations for the react compiler or the bug?
### How often does this bug happen?
Every time
### What version of React are you using?
19.2.0
### What version of React Compiler are you using?
1.0.0 | open | mimorisuzuko | 2025-11-17T07:22:49 | 2025-11-18T06:21:18 | null | [
"Type: Feature Request",
"Component: React Compiler"
] | null | 0 | {} | 6 | false | 2026-03-29T04:51:55.250152 | |||
facebook | react | 29,913 | I_kwDOAJy2Ks6MZP5_ | false | [React 19] React-Dom bundle increase 200kb | ## Summary
<!--
Please provide a CodeSandbox (https://codesandbox.io/s/new), a link to a
repository on GitHub, or provide a minimal code example that reproduces the
problem. You may provide a screenshot of the application if you think it is
relevant to your bug report. Here are some tips for providing a minimal
example: https://stackoverflow.com/help/mcve.
-->
## Using React 18
<img width="1153" alt="image" src="https://github.com/facebook/react/assets/58261676/d9d73fb7-f4bb-4cae-9a7d-19cf0bce7bca">
<img width="796" alt="image" src="https://github.com/facebook/react/assets/58261676/2d024945-39f4-4e0d-baec-60a56d791445">
## Useing React19
After Using React@canary, the react dom bundle changed to `react-dom-client.production.js` , which is 498kb!
<img width="823" alt="image" src="https://github.com/facebook/react/assets/58261676/e8e58714-fd19-4939-b188-3d5b616218fd">
<img width="1298" alt="image" src="https://github.com/facebook/react/assets/58261676/757efdde-1b84-4384-b44a-3e92c7737ff8">
| closed | completed | childrentime | 2024-06-16T05:09:55 | 2025-11-17T16:17:45 | 2024-06-17T02:26:36 | [
"React 19"
] | null | 0 | {} | 17 | false | 2026-03-29T04:51:55.295572 | ||
facebook | react | 30,994 | I_kwDOAJy2Ks6XBvQw | false | [React 19] [bug] SVG with dangerouslySetInnerHTML content does not trigger first click | ## Summary
Hi all,
Here is the scenario that we found out while testing with both the latest rc and the beta that works correctly with React 18.
We have an SVG element that does not trigger its first click if it is focusable (or positioned inside a focusable element) that changes some state on focus.
**Steps to reproduce:**
Open the Stackblitz example and open its console
Click 1 time on the triangle svg element
**Expected**:
'svg click' message is logged
**Current**:
no message is logged
(On the second and all next clicks the message is shown as expected - only the first click is not triggered)
**Here are the stackblitz examples where the issue can be observed:**
rc: https://stackblitz.com/edit/react-vsxt51-w3ktmp?file=app%2Fapp.tsx - not working
beta: https://stackblitz.com/edit/react-vsxt51-ssqptj?file=app%2Fapp.tsx - not working
**And here is how it is working in React 18:**
React 18: https://stackblitz.com/edit/react-vsxt51-xsg1yu?file=app%2Fapp.tsx - working
**Code**:
```
const App = () => {
const [focused, setFocused] = React.useState(false);
const handleFocus = () => {
setFocused(true);
};
return (
<svg
onFocus={handleFocus}
tabIndex={1}
onClick={() => {
console.log('svg click');
}}
viewBox="0 0 512 512"
dangerouslySetInnerHTML={{
__html: '<path d="M256 352 128 160h256z" />',
}}
></svg>
);
};
```
| open | zdravkov | 2024-09-18T13:52:05 | 2025-11-18T08:23:08 | null | [
"Type: Bug",
"React 19"
] | null | 0 | {
"+1": 12
} | 14 | false | 2026-03-29T04:51:55.358924 | |||
facebook | react | 32,362 | I_kwDOAJy2Ks6pwHPK | false | Bug: select field value resets to the first option after form submission with `useActionState` | When using a `<select>` field within a form that uses `useActionState`, the selected value resets to the first option after the form is submitted, even though the value should persist as `defaultValue` is set. This behavior occurs only with the `<select>` field; other form fields (e.g., text inputs) retain their values correctly after submission.
React version: 19
## Steps To Reproduce
1. Select 2 or 3 from the `select` field
2. Press the button to submit the form
Link to code example: https://codesandbox.io/p/sandbox/sweet-swirles-l8f2j6?file=%2Fsrc%2FApp.js&from-embed=
## The current behavior
After submission, the selected value resets to the first option (value='1').
## The expected behavior
The selected value should persist.
| closed | not_planned | joshua-ots | 2025-02-12T11:30:54 | 2025-11-18T14:05:32 | 2025-11-18T14:05:32 | [
"Status: Unconfirmed"
] | null | 0 | {
"+1": 25
} | 18 | false | 2026-03-29T04:51:55.402149 | ||
facebook | react | 10,657 | MDU6SXNzdWUyNTY0NDU2NzY= | false | Form restoration & React hydration | **Do you want to request a *feature* or report a *bug*?**
Bug
**Reproduction**
1. Open https://codesandbox.io/p/sandbox/form-restoration-react-hydration-cz1xgj?file=%2Fapp%2Fpage.js%3A2%2C1, see the source
2. Open https://cz1xgj-3000.csb.app/
3. Fill in the input, for example, you can type "foo"
4. Press submit
5. Undo the navigation
6. See how the "input value" is outdated, it doesn’t match what's inside the textbox.
https://github.com/facebook/react/assets/3165635/d29dd884-8896-4b59-8a55-16eb5145dfc7
**What is the current behavior?**
Let's say you start filling a form input type text or a select element with `foo`. Then you click on a link and press the back button. The browser back-forward cache (or maybe the form restauration logic) will kick in and change the initial value of the form elements. You will get the `foo` value back in the input. However, React doesn't account for it when the component is controlled.
For end-users, it means that they will try to submit a form because they see values in the inputs, but it won't work. They need to go back to each field and make a "fake" change so that React registers it.
This issue could also be worked around by not controlling the inputs, but I'm not sure how popular form libraries handle this.
**What is the expected behavior?**
I expect React to trigger an `onChange` event to replicate the actual form value.
At the fundamental level, there is a coordination issue between the browser and React.
**Which versions of React, and which browser / OS are affected by this issue? Did this work in previous versions of React?**
- React `@latest` (v18.2.0).
- Chrome `@latest` (v112.0)
**Details**
We have been using the following hack on Material UI:
```js
componentDidMount() {
// Fix SSR issue with the go back feature of the browsers.
// Let's say you start filling the input with "foo", you change the page then after comes back.
// The browser will reset the input value to "foo", but we also need to tell React about it.
this.handleChange({
target: this.input,
});
}
```
But it comes with issues: https://github.com/mui/material-ui/pull/8110.
So instead, we tried simulating a change event [as suggested in Stack Overflow](https://stackoverflow.com/questions/23892547/what-is-the-best-way-to-trigger-onchange-event-in-react-js) but it doesn't work anymore
```js
var event = new Event('input', { bubbles: true });
element.dispatchEvent(event);
```
So for now, we are going to disable the cache with `<form autoComplete="off">`.
| open | oliviertassinari | 2017-09-09T15:23:08 | 2025-11-18T14:07:25 | null | [
"Type: Bug",
"Type: Feature Request",
"Component: DOM"
] | null | 0 | {
"+1": 2
} | 4 | false | 2026-03-29T04:51:55.684856 | |||
facebook | react | 33,967 | I_kwDOAJy2Ks7B8jya | false | Bug: Exception thrown by Error Boundary after production build in React 19, not visible in UI but reported via monitoring | After deploying my application using React 19 to production, I intermittently receive error notifications triggered by my custom error boundary (i.e., componentDidCatch in a class-based ErrorBoundary component).
The strange behavior is:
• These exceptions do not appear in the UI — the app works as expected visually.
• The error seems to occur only after the app has been used for a few days post-deployment.
• I am able to catch the exception via monitoring tools (e.g., Sentry, LogRocket, etc.), but cannot reproduce it or observe any direct malfunction in the UI.
• This was not happening with React 18, and only began after migrating to React 19.
⸻
Reproduction steps
Unfortunately, the issue is hard to reproduce consistently, but here’s the general flow:
1. Deploy app to production using React 19.
2. Use the app normally for a first time.
3. Receive exception reports from componentDidCatch (without seeing any UI impact).
4. The component tree that fails is not removed or visibly broken.
⸻
Expected behavior
• If an exception is thrown and caught by an error boundary, I expect it to show a fallback UI (or at least interrupt rendering).
• Silent failures with no visible issues but background error boundary triggers are unexpected.
⸻
Actual behavior
• Exception is thrown in production and caught by componentDidCatch.
• No fallback UI is rendered.
• No visual issues seen by users.
• Only monitoring tools detect the issue.
⸻
Additional context
• Using ErrorBoundary class component.
• Exception is only caught in production builds (built with vite build / webpack).
• Could this be related to React 19’s new compiler or effects handling? | closed | not_planned | iamamitshrma | 2025-07-22T19:45:44 | 2025-10-28T18:08:06 | 2025-10-28T18:08:06 | [
"Status: Unconfirmed",
"Resolution: Stale"
] | null | 0 | {} | 3 | false | 2026-03-29T04:51:50.294546 | ||
facebook | react | 32,960 | I_kwDOAJy2Ks6zGQrF | false | Feature Request: add useEffect callback argument | This is not a bug, but I didn't find any other place to ask for new features. Sorry if it bothers you.
It would be really useful to pass an argument to a useEffect callback and get an array of the dependencies that triggered the effect, eventualy with their previous values. Something like that :
```
useEffect(prev => {
console.log(prev)
}, [dep1, dep2, dep3]
//=> [ { name: 'dep1', prevValue: <value> }, { name: 'dep3', prevValue: <value>} ]
```
If there is a better existing way to achieve the same thing, please tell me. | closed | not_planned | cgx-system | 2025-04-18T11:32:48 | 2025-10-28T23:05:05 | 2025-10-28T23:05:05 | [
"Status: Unconfirmed",
"Resolution: Stale"
] | null | 0 | {
"hooray": 1
} | 7 | false | 2026-03-29T04:51:50.350149 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.