status
stringclasses 1
value | repo_name
stringlengths 9
24
| repo_url
stringlengths 28
43
| issue_id
int64 1
104k
| updated_files
stringlengths 8
1.76k
| title
stringlengths 4
369
| body
stringlengths 0
254k
⌀ | issue_url
stringlengths 37
56
| pull_url
stringlengths 37
54
| before_fix_sha
stringlengths 40
40
| after_fix_sha
stringlengths 40
40
| report_datetime
timestamp[ns, tz=UTC] | language
stringclasses 5
values | commit_datetime
timestamp[us, tz=UTC] |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
closed
|
facebook/react
|
https://github.com/facebook/react
| 12,984 |
["packages/react-dom/src/server/ReactPartialRenderer.js"]
|
nested contexts don't unwind correctly when server side rendering
|
**Do you want to request a *feature* or report a *bug*?**
bug
**What is the current behavior?**
When different contexts are nested, as in they have different types, they do not unwind correctly when rendering via `renderToString()`.
Here is a snippet that demonstrates the issue:
```
import React from 'react';
import { renderToString } from 'react-dom/server';
const valueA = { letter: 'A' };
const valueB = { letter: 'B' };
const LetterContext = React.createContext(valueA);
const NumberContext = React.createContext(undefined);
const html = renderToString((
<LetterContext.Provider value={valueA}>
<NumberContext.Provider>
<LetterContext.Consumer>
{letterValue => (
<div>
{letterValue.letter}
<LetterContext.Provider value={valueB}>
<LetterContext.Consumer>
{innerValue => <div>{innerValue.letter}</div>}
</LetterContext.Consumer>
</LetterContext.Provider>
</div>
)}
</LetterContext.Consumer>
<LetterContext.Consumer>
{value => <div>{value.letter}</div>}
</LetterContext.Consumer>
</NumberContext.Provider>
</LetterContext.Provider>
));
console.log(html);
```
This results in:
```
value.letter
^
TypeError: Cannot read property 'letter' of undefined
at Object.children (/Users/ericsoderberg/Documents/git/grommet2/grommet-icons-site/tools/example.js:25:31)
at ReactDOMServerRenderer.render (/Users/ericsoderberg/Documents/git/grommet2/grommet-icons-site/node_modules/react-dom/cjs/react-dom-server.node.development.js:2462:55)
at ReactDOMServerRenderer.read (/Users/ericsoderberg/Documents/git/grommet2/grommet-icons-site/node_modules/react-dom/cjs/react-dom-server.node.development.js:2325:19)
at renderToString (/Users/ericsoderberg/Documents/git/grommet2/grommet-icons-site/node_modules/react-dom/cjs/react-dom-server.node.development.js:2697:25)
at Object.<anonymous> (/Users/ericsoderberg/Documents/git/grommet2/grommet-icons-site/tools/example.js:9:14)
at Module._compile (internal/modules/cjs/loader.js:678:30)
at loader (/Users/ericsoderberg/Documents/git/grommet2/grommet-icons-site/node_modules/babel-register/lib/node.js:144:5)
at Object.require.extensions.(anonymous function) [as .js] (/Users/ericsoderberg/Documents/git/grommet2/grommet-icons-site/node_modules/babel-register/lib/node.js:154:7)
at Module.load (internal/modules/cjs/loader.js:589:32)
at tryModuleLoad (internal/modules/cjs/loader.js:528:12)
```
**What is the expected behavior?**
When context providers are nested, they should correctly unwind such that the contexts return the correct value at each level. No exception should be generated and the output should be:
`<div>A<div>B</div></div><div>A</div>`
**Which versions of React, and which browser / OS are affected by this issue? Did this work in previous versions of React?**
This was uncovered using version 16.4.0 of `react-dom` on OSX using node v10.1.0.
This only shows up when using `renderToString()`. It works as expected in the browser.
** Note **
I wonder if this might be due to the following snippet of `react-dom-server.node.development.js`. Where the previous provider added to the stack is assumed to have the same type as the provider being popped.
```
var context = provider.type._context;
if (this.providerIndex < 0) {
context._currentValue = context._defaultValue;
} else {
// We assume this type is correct because of the index check above.
var previousProvider = this.providerStack[this.providerIndex];
context._currentValue = previousProvider.props.value;
}
```
I am investigating further and hope to provide a pull request soon.
|
https://github.com/facebook/react/issues/12984
|
https://github.com/facebook/react/pull/13019
|
30bc8ef79295c71f5a4fc459c5bbd4b271f6f390
|
ec60457bcd22d06c404a44839c14beda88081ac7
| 2018-06-05T21:00:28Z |
javascript
| 2018-06-11T19:52:39Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 12,964 |
["packages/react-dom/src/__tests__/ReactDOMFiber-test.js", "packages/react-reconciler/src/ReactChildFiber.js"]
|
Returning an empty fragment throws a confusing error
|
https://codesandbox.io/s/40j50607y0
Rendering a component like:
```js
const Foo = () => <React.Fragment />
```
Throws the error:
> Foo(...): Nothing was returned from render. This usually means a return statement is missing. Or, to render nothing, return null.
The error is confusing because clearly `Foo` _is_ returning a value. I don't know if we want to support rendering empty fragments, but in any case this error is confusing.
If this isn't allowed it should explicitly call out that a component cannot return a fragment with no children.
cc @gaearon
|
https://github.com/facebook/react/issues/12964
|
https://github.com/facebook/react/pull/12966
|
4ac6f133aff78e7616c7427167f0d43a18441b94
|
d480782c4162431d06c077ebf8fdf6b8ba7896ef
| 2018-06-01T21:11:00Z |
javascript
| 2018-06-11T13:43:30Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 12,930 |
["packages/react-scheduler/src/ReactScheduler.js", "packages/shared/requestAnimationFrameForReact.js"]
|
Broken `fiber-triangle-demo` and `schedule` fixtures
|
**Do you want to request a *feature* or report a *bug*?**
Bug
**What is the current behavior?**
These two fixtures are both broken due to a reference to `global` in the bundle they reference.
This was introduced in a PR I just landed, so I'm fixing it. See https://github.com/facebook/react/pull/12900
**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:**
Open either fixture in Firefox, note the error in the console and the page doesn't update.
**What is the expected behavior?**
They should have React and/or the schedule module running JS on the page.
**Which versions of React, and which browser / OS are affected by this issue? Did this work in previous versions of React?**
Just on master.
|
https://github.com/facebook/react/issues/12930
|
https://github.com/facebook/react/pull/12931
|
ff724d3c286a1753723ea71e8c046498ed1aac98
|
79a740c6e32ca300d4e7ff55ab06de172d4237fd
| 2018-05-29T23:01:14Z |
javascript
| 2018-05-30T00:54:38Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 12,926 |
["packages/react-dom/src/__tests__/ReactDOMServerIntegrationForms-test.js", "packages/react-dom/src/client/ReactDOMFiberComponent.js", "packages/react-dom/src/client/ReactDOMFiberInput.js"]
|
`onChange` event doesn't trigger when direclty clicking on the maximum `input[type=range]` value
|
<!--
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*?**
*bug*
**What is the current behavior?**
When you directly after the loading choose the maximum value on the input type range, the event isn't fire with an onChange. The event is fire when it's not the maximum value of the input.
Sandbox: https://codesandbox.io/s/7kv59yj360
Click on the maximum (right) of the input of type range.
There is no message on the console and the displayed value doesn't change as expected
Click on a other value on the input, a message is displayed in the console and the value change
When you click on the last value it's now displayed.
**What is the expected behavior?**
An event should trigger when you directly click on the maximum value of the input of type range.
**Which versions of React, and which browser / OS are affected by this issue? Did this work in previous versions of React?**
Testing on macOS High Sierra, Firefox 60.0.1, Safari 11.1, Chrome 66.0.3359.181, react 16.2.0
To get around this behaviour, I had to change onChange by onInput but add a blanck onChange to still get the drag behaviour on Safari on Iphone (see this [CodeSandBox ](https://codesandbox.io/s/w0x0jzjvw5))
|
https://github.com/facebook/react/issues/12926
|
https://github.com/facebook/react/pull/12939
|
65ab53694f6edf4437034d5644a602dea162e249
|
36546b5137e9012ebdc62fc9ec11e3518c9e0aab
| 2018-05-29T16:00:41Z |
javascript
| 2018-05-31T21:23:26Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 12,872 |
["packages/react-dom/src/__tests__/ReactDOMInput-test.js", "packages/react-dom/src/client/ReactDOMFiberInput.js"]
|
Submit/Reset inputs lose text when value=undefined.
|
<!--
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*?**
Bug.
**What is the current behavior?**
Creating an `<input type='submit'` with `value={undefined}` results in a button with no text. https://codesandbox.io/s/nn7p94y3rl
The same behaviour can be seen with `type='reset'`
I have a PR open with a proposed fix. https://github.com/facebook/react/pull/12780
**What is the expected behavior?**
The input should use the browser's default value for the input (Submit for en, etc.).
**Which versions of React, and which browser / OS are affected by this issue? Did this work in previous versions of React?**
React 16.3.2. Chrome 66.0.3359.139. macOS 10.13.4. Works in React 15 (unsure of earlier versions)
|
https://github.com/facebook/react/issues/12872
|
https://github.com/facebook/react/pull/12780
|
8862172fa3d9f30ea6a7c1072bdae068e9794856
|
9832a1b6d5df06c3d2a54febe2ab3de43ace02fc
| 2018-05-21T01:28:08Z |
javascript
| 2018-08-16T14:19:03Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 12,870 |
["packages/react-reconciler/README.md"]
|
Broken link for react-reconciler/README.md Practical Examples for React Native
|
On [README.md](https://github.com/facebook/react/blob/master/packages/react-reconciler/README.md), the link for React Native Practical Example is broken.
|
https://github.com/facebook/react/issues/12870
|
https://github.com/facebook/react/pull/12871
|
9bed4a6aee557e2f7425705c64e49369125cceeb
|
d7b9b4921b9f5bd12500df9b9f164caa4bcd75eb
| 2018-05-20T01:02:09Z |
javascript
| 2018-05-20T11:01:00Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 12,765 |
["packages/react-dom/src/__tests__/ReactDOMInput-test.js", "packages/react-dom/src/client/ReactDOMFiberInput.js"]
|
defaultChecked prop sets only checked attribute
|
#### Type
`inconsistency` for common use cases.
`bug` for projects relying on correct DOM emissions.
#### Description
It's unclear to me if the following is a known issue or not.
Take in consideration the snippet below:
```jsx
<input type="checkbox" defaultChecked={true} />
```
#### Current Behavior
Setting the `defaultValue` prop on inputs of other types sets both the `value` and `defaultValue` attribute on the DOM element.
When setting the `defaultChecked` prop on inputs of type `checkbox` only the `checked` attribute is correctly set, leaving the `defaultChecked` attribute unchanged.
#### Expected Behaviour
Like mentioned above, the expected behavior would be for the `defaultChecked` prop to set both the `checked` and `defaultChecked` attributes on the corresponding DOM node (like it's the case with the `defaultValue` prop).
|
https://github.com/facebook/react/issues/12765
|
https://github.com/facebook/react/pull/13114
|
07fefe33310849df81836c1c2e773b7d50c3ec2b
|
85fe4ddce7ae4d8c93a8107db9e7e65a0631baeb
| 2018-05-08T14:15:22Z |
javascript
| 2018-07-04T20:00:42Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 12,694 |
["packages/react-dom/src/__tests__/DOMPropertyOperations-test.js", "packages/react-dom/src/shared/DOMProperty.js"]
|
Size prop is not attached to the DOM
|
**Do you want to request a *feature* or report a *bug*?**
Bug
**What is the current behavior?**
The size prop can only be a number and I think this should be true for inputs and selects but not necessarily true for the rest of HTML tags or web components.
**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:**
Sorry, it's a CodePen. This example uses a framework called OnsenUI that internally uses web components. In the OnsenUI framework there is an icon web component that transforms every size prop value into a string which is not attached to the DOM.
I know that the example has more dependencies than React, BUT there is also a simple div (as simple as this `<div size="40px" />`) which is also affected by this issue.
https://codepen.io/airamrguez/pen/bMEgEP
**What is the expected behavior?**
The size prop should be added to the DOM when it is not a number on HTML tags that aren't inputs and selects.
**Which versions of React, and which browser / OS are affected by this issue? Did this work in previous versions of React?**
The affected versions are >= 16.3
It works ok in versions <= 16.2
Browsers: All
|
https://github.com/facebook/react/issues/12694
|
https://github.com/facebook/react/pull/12702
|
045d4f166db6023550b76da6a83fe595a6cdf2f2
|
dcc854bcc3c940ca583565ce25200ca618c05bf0
| 2018-04-26T06:50:06Z |
javascript
| 2018-04-28T19:52:30Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 12,688 |
["packages/react-reconciler/src/ReactFiberBeginWork.js", "packages/react/src/__tests__/forwardRef-test.internal.js"]
|
Unexpected behavior when use React.forwardRef with React.createContext
|
<!--
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*?**
bug
**What is the current behavior?**
example: https://codesandbox.io/s/04393o3k6w
If the `LogProps` is child of `M.Provider`, and `M.Consumer` is used in its children.
When the `M.Provider` is updating, even if the `LogProps` is wrapped in a component that never updates, it can still be triggered by the callback of `React.forwardRef`, and updates the `LogProps`.
**What is the expected behavior?**
the `LogProps` shouldn't update when the root component had updated, whether or not `M.Consumer` is used in its children;
**Which versions of React, and which browser / OS are affected by this issue? Did this work in previous versions of React?**
it's used [email protected]
|
https://github.com/facebook/react/issues/12688
|
https://github.com/facebook/react/pull/12690
|
ec57d2994156ba5fcf794100968764bf25206629
|
d883d59863483a1fb9f3dd0455022e4c6e1a41d5
| 2018-04-25T14:16:43Z |
javascript
| 2018-04-26T18:47:34Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 12,686 |
["packages/react-reconciler/src/ReactFiberBeginWork.js", "packages/react-reconciler/src/__tests__/ReactNewContext-test.internal.js"]
|
Strange beahvior React + new Context API
|
I think it is a bug.
Git with example - https://github.com/axules/react-context-research
> npm i
npm start
go to http://localhost:3003
**Dependencies**
React v16.3.2
Chrome v65.0.3325.181
### Components
Context - OrderDetailsContext
Root component - OrdersView
Simple component - OrdersViewRow
Subscribed component - OrderDetailsView
### Problem
Application have one dynamic context (OrderDetailsContext) and 3 components: one root component (OrdersView) and 2 children components (OrdersViewRow and OrderDetailsView). Only one component (OrderDetailsView) subscribed to dynamic context and have to be refreshed once context data is changed. But with them we have problem.
When context is changed the subscribed component is not refreshed and have old version of context data.
#### actual
OrderDetailsView do not update once context is updated.
#### expected
OrderDetailsView will update once context is updated.
### First solution
Go to OrdersView and move OrderDetailsView to the top of render function, then subscribed component will be updated when context is changed.
### Second strange solution
Go to separeted component (OrdersViewRow) and just remove `<td>Remove this td and I will work!</td>`, then subscribed component will be updated when context is changed.
### Video
https://youtu.be/WET0a2hApwM
### Reproduce
1. Clone https://github.com/axules/react-context-research.git
1. npm i
1. npm start
1. Open http://localhost:3003 in Chrome
1. Open dev tools in Chrome
1. Click on View button
#### expected
* 'Loading...' text should be visibled and hidden after.
* Text should be updated and new number should be displayed.
* console log should be
#### actual
* 'Loading...' text is not visibled.
* Text is not changed.
* console log is
|
https://github.com/facebook/react/issues/12686
|
https://github.com/facebook/react/pull/12708
|
7c3932857195d0cfba2379875b65f1bd1fae39fd
|
045d4f166db6023550b76da6a83fe595a6cdf2f2
| 2018-04-25T06:19:03Z |
javascript
| 2018-04-28T00:52:48Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 12,670 |
["packages/react-dom/src/__tests__/ReactComponentLifeCycle-test.js", "packages/react-dom/src/__tests__/ReactDOMServerLifecycles-test.js", "packages/react-dom/src/server/ReactPartialRenderer.js", "packages/react-reconciler/src/ReactFiberClassComponent.js", "packages/react/src/__tests__/ReactCoffeeScriptClass-test.coffee", "packages/react/src/__tests__/ReactES6Class-test.js", "packages/react/src/__tests__/ReactTypeScriptClass-test.ts", "packages/react/src/__tests__/createReactClassIntegration-test.js"]
|
Improve warning message for failure to initialize `state` when using `getDerivedStateFromProps`
|
**Do you want to request a *feature* or report a *bug*?**
bug
**What is the current behavior?**
The warning for a failure to initialize state when using `getDerivedStateFromProps` is not as clear as it could be.
This component produces the warning:
> MyComponent: Did not properly initialize state during construction. Expected state to be an object, but it was undefined.`
```js
class MyComponent extends React.Component {
static getDerivedStateFromProps(nextProps, prevState) {
return null;
}
render() {
return <span>Foo</span>
}
}
```
The warning can be resolved by adding `state = {}` either in the class definition or the constructor, or by not using `getDerivedStateFromProps`.
See https://github.com/reactjs/reactjs.org/issues/796 for more detail on why this warning message can be counter-intuitive.
Since the [warning is specific](https://github.com/facebook/react/blob/b548b3cd640dbd515f5d67dafc0216bb7ee0d796/packages/react-reconciler/src/ReactFiberClassComponent.js#L509-L510) to the use of `getDerivedStateFromProps` and not triggered by any other state or lifecycle methods, and state initialization isn't required for other state methods like `this.setState`, it would make sense to me to mention `getDerivedStateFromProps` explicitly in the warning.
Here's a proposal for a new warning message:
> MyComponent: Component state must be initialized when using getDerivedStateFromProps. Expected state to be an object, but it was undefined.
**Which versions of React, and which browser / OS are affected by this issue? Did this work in previous versions of React?**
React >= 16.3.0
|
https://github.com/facebook/react/issues/12670
|
https://github.com/facebook/react/pull/13317
|
fa824d092195713bf85d56cbda8bad976bbc1900
|
b179bae0ae3a9d80bfbe440c8861786b7f928ce1
| 2018-04-23T15:29:28Z |
javascript
| 2018-08-03T15:09:57Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 12,572 |
["packages/react-test-renderer/package.json", "scripts/rollup/bundles.js"]
|
Feature Request: UMD version of react-test-renderer
|
react and react-dom have both cjs and umd versions available.
Is it possible to create a umd version of react-test-renderer as well?
|
https://github.com/facebook/react/issues/12572
|
https://github.com/facebook/react/pull/12594
|
3eae866e03a96c4f46e257cba73ca158b049ab05
|
b8461524db6d3e016fabf001ad8fa086b4918ef9
| 2018-04-08T02:37:39Z |
javascript
| 2018-04-10T14:49:19Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 12,551 |
["packages/react-reconciler/src/ReactFiberBeginWork.js", "packages/react-reconciler/src/__tests__/ReactNewContext-test.internal.js"]
|
New Context Provider may block Old context propagation if children are constant
|
<!--
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*?**
Bug
**What is the current behavior?**
It seems that, if the children of a new-style `React.createContext()` context `Provider` are constant, the `Provider` can block updates from old-style `this.context` context providers from propagating to `this.context` consumers.
This sandbox demonstrates the issue. Clicking the button with a number will correctly increment the `Root`'s `state` and `context`, but the update is only propagated to the `Child3`'s `context` (and its button) when the "Colors!" button is clicked, as it causes an update to the `value` of the new-style `Provider`:
https://codesandbox.io/s/ol4lpokpjy
<details>
<summary>Copy of the source code in the sandbox</summary>
```jsx
import PropTypes from "prop-types";
import React from "react";
import ReactDOM from "react-dom";
class Root extends React.Component {
constructor(props: {}) {
super(props);
this.state = {
count: 0
};
this.countUp = this.countUp.bind(this);
}
getChildContext() {
return {
...this.context,
count: this.state.count,
countUp: this.countUp
};
}
render() {
return this.props.children;
}
countUp() {
this.setState(({ count }) => ({ count: count + 1 }));
}
}
Root.childContextTypes = {
count: PropTypes.number.isRequired,
countUp: PropTypes.func.isRequired
};
const ctx = React.createContext();
class Child1 extends React.Component {
constructor(props: { onClick(): void }) {
super(props);
this.state = {
color: randomHexColor(),
newColor: this.newColor.bind(this)
};
}
render() {
return (
<ctx.Provider value={this.state}>{this.props.children}</ctx.Provider>
);
}
newColor() {
const color = randomHexColor();
this.setState(() => ({ color }));
}
}
function randomHexColor() {
const colorStr = Math.floor(Math.random() * (Math.pow(2, 24) - 1)).toString(
16
);
return "#000000".slice(0, -colorStr.length) + colorStr;
}
class Child2 extends React.Component {
render() {
return (
<ctx.Consumer>
{ctx => (
<React.Fragment>
<Child3 color={ctx.color} />
<button onClick={ctx.newColor}>Colors!</button>
</React.Fragment>
)}
</ctx.Consumer>
);
}
}
class Child3 extends React.Component {
render() {
return (
<button
style={{ color: this.props.color }}
onClick={this.context.countUp}
>
{this.context.count}
</button>
);
}
}
Child3.contextTypes = {
count: PropTypes.number.isRequired,
countUp: PropTypes.func.isRequired
};
ReactDOM.render(
<Root>
<Child1>
<Child2 />
</Child1>
</Root>,
document.getElementById("root")
);
```
</details>
**What is the expected behavior?**
Both old-style and new-style context updates should coexist.
**Which versions of React, and which browser / OS are affected by this issue? Did this work in previous versions of React?**
16.3.1; also broken in 16.3.0.
---
This seems to only happen if the `children` of the `Provider` are constant, which is what happens when the children are provided on the first and only `ReactDOM.render` call. If `Child1` is updated to directly use `<Child2/>` instead of `{this.props.children}`, the problem does not happen.
This can also be a problem when using a production optimization that hoists constant elements outside the Component if the specified children are constant, which would even defeat the fix/workaround for the example above.
|
https://github.com/facebook/react/issues/12551
|
https://github.com/facebook/react/pull/12586
|
d883d59863483a1fb9f3dd0455022e4c6e1a41d5
|
7c3932857195d0cfba2379875b65f1bd1fae39fd
| 2018-04-05T09:18:08Z |
javascript
| 2018-04-26T19:59:17Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 12,540 |
["packages/react-reconciler/src/ReactStrictModeWarnings.js"]
|
react-dom development bundle uses Array.from
|
<!--
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*?**
Bug (?)
**What is the current behavior?**
Currently, the 16.3.x versions of `react-dom` ship [`react-dom.development.js`](https://unpkg.com/[email protected]/cjs/react-dom.development.js) containing usage of `Array.from` which is not available in certain React-supported browsers like IE11. It causes breakage if those particular code paths are taken in such a browser (I ran into it using `<StrictMode>` in IE11).
**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:**
Reproduction (should see "Hello, world!" but IE11 will crash and show nothing):
https://codepen.io/anon/pen/zWJBrO?editors=0010
**What is the expected behavior?**
`react-dom.development.js` would not use `Array.from` or document that a polyfill is required.
**Which versions of React, and which browser / OS are affected by this issue? Did this work in previous versions of React?**
`react` and `react-dom`: 16.3.0-alpha.0 through 16.3.1
Internet Explorer 11
|
https://github.com/facebook/react/issues/12540
|
https://github.com/facebook/react/pull/12546
|
d328e362e86a6af4a0664e004b8f97f18ce972c8
|
8ec0e4a99df76c0ff1779cac4f2eaaaf35a6b5bb
| 2018-04-04T17:01:55Z |
javascript
| 2018-04-04T20:54:27Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 12,506 |
["packages/react-dom/src/events/getEventTarget.js"]
|
Possible incorrect event.target on number inputs in IE9?
|
Since React 16.0.0, it looks like there might be a case where `event.target` on a change event in IE9 reports as the window, but I can't reproduce it outside of the DOM test fixtures.
*Steps to reproduce*
1. Open this build of the fixtures: http://react-ie-9-target-issues.surge.sh in IE9
2. Enter text into an input
3. Observe that the `event.target.value` is the window, as recorded by the fixture
I can not reproduce this using the following CodePen:
https://codepen.io/nhunzaker/pen/dmeoxJ?editors=1010
Viewable in IE9 here: https://s.codepen.io/nhunzaker/debug/dmeoxJ/PNrvYLevqQbM
Here's a diff of my changes with master:
https://github.com/facebook/react/compare/master...nhunzaker:ie9-target-issue
Any idea what's going on? Here's what I think I need to check next:
- [x] Differences in rAF polyfill
- [x] Differences in map/set polyfill
_This list will probably grow_
|
https://github.com/facebook/react/issues/12506
|
https://github.com/facebook/react/pull/12976
|
23be4102df273d6ee6d9be42c6d65885d0a914b1
|
4ac6f133aff78e7616c7427167f0d43a18441b94
| 2018-03-31T14:45:51Z |
javascript
| 2018-06-11T13:35:42Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 12,502 |
["packages/react-reconciler/src/ReactFiber.js", "packages/react-reconciler/src/ReactFiberScheduler.js", "packages/react-reconciler/src/__tests__/ReactIncrementalErrorReplay-test.internal.js"]
|
Possible regression in dev mode in v16.3
|
[This example](https://codesandbox.io/s/1zn2o513pq) presents a huge performance drop between React 16.2 and React 16.3.
In investigated it and noticed several things:
- Perfomance drop is only visible using development bundle
- This is not related to styled-components (a simple div presents the same problem)
- I tried to investigate in performance tab of Chrome dev tools and I noticed that I get React measures even without "?react_perf"
- React Tree Reconciliation seems very long, do not know why
It is not a big problem for me, but it looks weird, so I decided to [share it on Twitter](https://twitter.com/neoziro/status/979756579296817154), then @gaearon [invited me to submit an issue about it](https://twitter.com/dan_abramov/status/980036013518139398).
|
https://github.com/facebook/react/issues/12502
|
https://github.com/facebook/react/pull/12510
|
0c80977061ba576cee9ae0891245be233929d2ed
|
59dac9d7a6a2f0b66003cf717d71b5587265423f
| 2018-03-31T11:22:27Z |
javascript
| 2018-04-01T18:10:37Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 12,500 |
["packages/react/src/ReactElementValidator.js"]
|
Warning in IE 11 when using React.Fragment
|
**Do you want to request a *feature* or report a *bug*?**
bug
**What is the current behavior?**
With React 16.3.0, when using `<React.Fragment>` IE 11 gives the following warning:
```
Warning: Invalid prop `children` supplied to `React.Fragment`. React.Fragment can only have `key` and `children` props.
```
Steps to reproduce:
1. Use `create-react-app` to create a new React app.
2. Add `<React.Fragment>` to `App.js`. For example:
```
import React, { Component } from "react";
import logo from "./logo.svg";
import "./App.css";
class App extends Component {
render() {
return (
<div className="App">
<React.Fragment>
<header className="App-header">
<img src={logo} className="App-logo" alt="logo" />
<h1 className="App-title">Welcome to React</h1>
</header>
<p className="App-intro">
To get started, edit <code>src/App.js</code> and save to reload.
</p>
</React.Fragment>
</div>
);
}
}
export default App;
```
3. Open the app in IE 11.
4. Open IE dev tools and refresh the browser.
5. You should see the above mentioned warning message in the console window.
No warnings with Chrome, Firefox, and Edge.
**What is the expected behavior?**
There should be no warnings shown.
**Which versions of React, and which browser / OS are affected by this issue? Did this work in previous versions of React?**
React: 16.3.0
Browser: IE 11
OS: Windows 10
|
https://github.com/facebook/react/issues/12500
|
https://github.com/facebook/react/pull/12504
|
fa8e67893fca1b3902637129972032bca248a584
|
0c80977061ba576cee9ae0891245be233929d2ed
| 2018-03-31T11:00:44Z |
javascript
| 2018-04-01T00:14:36Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 12,481 |
["fixtures/attribute-behavior/AttributeTableSnapshot.md", "packages/react-dom/src/shared/DOMProperty.js"]
|
[SVG] Enable focusable to accept boolean values
|
Moved discussion from https://github.com/facebook/react/issues/6212.
## Situation
The [`focusable` attribute](https://www.w3.org/TR/SVGTiny12/interact.html#focusable-attr) from the SVG specifications is an [enumerated attribute](https://html.spec.whatwg.org/#keywords-and-enumerated-attributes) accepting values `"true"`, `"false"` and `"auto"`. Because it is technically not a boolean attribute (although it certainly somehow looks like it), React expects the value to be passed as a string. See the following example:
```diff
-<svg focusable>I should be focusable</svg>
-<svg focusable={true}>I should be focusable</svg>
+<svg focusable='true'>I should be focusable</svg>
```
The thing is, the `focusable` attribute is often used in conjunction with elements from the ARIA specification, in which attributes are booleans and not enumerated attributes with `"true"` and `"false"` values. The [`aria-hidden` attribute](https://www.w3.org/TR/wai-aria-1.1/#aria-hidden) is a good example of that.
For instance, [following a good practice for icon-buttons](https://fvsch.com/code/svg-icons/#section-html):
```html
<button type="button">
<svg aria-hidden="true" focusable="false">
<use xlink:href="#icon-play"></use>
</svg>
<span class="access-label">Start playback</span>
</button>
```
From an authoring perspective, the above snippet would likely be written like this in JSX:
```jsx
<button type='button'>
<Icon icon='play' aria-hidden={true} focusable={false} />
<span class='access-label'>Start playback</span>
</button>
```
The problem is that `focusable` **cannot** be authored as a boolean, otherwise it will **not** be printed out in the DOM. On the other hand, `aria-hidden` is perfectly fine being written as a boolean at it gets coerced by React.
## Proposal
Given the default value for the `focusable` attribute is `"auto"`, this is very likely this attribute gets authored to change its value to `true` or `false`. In that regard, it is confusing that it has to be specified as a string, when other attributes accepting booleans can be authored as such.
The suggestion would be to make it possible for `focusable` to be specified as either a boolean or a string, like other similar attributes. In other words, all the following should work:
```jsx
<svg focusable>I should be focusable</svg>
<svg focusable={true}>I should be focusable</svg>
<svg focusable='true'>I should be focusable</svg>
<svg focusable={false}>I should not be focusable</svg>
<svg focusable='false'>I should not be focusable</svg>
<svg focusable='auto'>I should be focusable</svg>
```
From an authoring perspective, I believe this would be the most straightforward and less confusing.
|
https://github.com/facebook/react/issues/12481
|
https://github.com/facebook/react/pull/13339
|
a66217b5716a5a32e6fb80bd1774610b05c4be12
|
3cfab14b96a55d53ba5784a9915b8c1660da08be
| 2018-03-29T08:35:33Z |
javascript
| 2018-08-07T18:39:56Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 12,474 |
["packages/react-dom/src/__tests__/ReactDOMRoot-test.internal.js", "packages/react-reconciler/src/ReactFiberScheduler.js"]
|
Uncaught Error: Should be working on a root. This error is likely caused by a bug in React.
|
Got this debugging another error. Filing here for reference. Original code: https://phabricator.intern.facebook.com/P59317705
|
https://github.com/facebook/react/issues/12474
|
https://github.com/facebook/react/pull/12480
|
268a3f60dfe67c4f6439fc37b569a2d81c81a53a
|
c44665e83278becfe7a3afdf788789536d63387b
| 2018-03-28T15:03:46Z |
javascript
| 2018-03-29T01:18:09Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 12,432 |
["packages/react-test-renderer/src/ReactShallowRenderer.js", "packages/react-test-renderer/src/__tests__/ReactShallowRenderer-test.internal.js", "packages/react-test-renderer/src/__tests__/ReactShallowRenderer-test.js"]
|
Remove DEV-only warnings from shallow renderer
|
For consideration.
These warnings are already covered by renderers based on react-reconciler. Maybe we don't gain that much by mirroring them in the shallow renderer? Maybe it's not worth the added complexity? Particularly with more complicated warnings like the ones being added for #12419 and #12404.
|
https://github.com/facebook/react/issues/12432
|
https://github.com/facebook/react/pull/12433
|
c1308adb4bbe9e5513bbc4499e0696d97e9c897f
|
8c20615b06d22e023ff54a30e2602f0409876441
| 2018-03-22T18:08:16Z |
javascript
| 2018-03-22T18:32:37Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 12,428 |
[".eslintrc.js", "packages/react-dom/src/__tests__/ReactCompositeComponent-test.js", "packages/react-dom/src/__tests__/ReactDOMTextarea-test.js", "packages/react-dom/src/__tests__/ReactServerRendering-test.js", "scripts/eslint-rules/__tests__/no-to-warn-dev-within-to-throw-test.internal.js", "scripts/eslint-rules/index.js", "scripts/eslint-rules/no-to-warn-dev-within-to-throw.js", "scripts/jest/matchers/__tests__/toWarnDev-test.js"]
|
We don't assert on warnings for tests that throw
|
I think this shouldn't pass:
```js
expect(() => {
expect(() => {
throw new Error('haha') // no warning
}).toWarnDev(
'nope',
);
}).toThrow('haha');
```
but it does.
|
https://github.com/facebook/react/issues/12428
|
https://github.com/facebook/react/pull/12457
|
026aa9c97813ee8e9886e2c9035a332f93c38720
|
e106b8c44f9a81058f250c7ee37b61ae0094455e
| 2018-03-22T16:57:57Z |
javascript
| 2018-08-21T14:43:02Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 12,427 |
["scripts/jest/matchers/toWarnDev.js"]
|
jestDiff() call in toWarnDev() returns null, making test failure confusing
|
<img width="953" alt="screen shot 2018-03-22 at 4 47 40 pm" src="https://user-images.githubusercontent.com/810438/37785026-037d0788-2df1-11e8-9cbd-5b9e2b2d10db.png">
Caused by this line:
https://github.com/facebook/react/blob/40fa6160534e2bc67b0c43f0b32dcdd1d214068b/scripts/jest/matchers/toWarnDev.js#L52
Calling `jestDiff` with two different strings returns `null`.
|
https://github.com/facebook/react/issues/12427
|
https://github.com/facebook/react/pull/12456
|
dadafd6bd802082b990e447032cb82574bcdc118
|
e9ba8ec8663071ebe9467c13576b3c18ad4e1b58
| 2018-03-22T16:50:03Z |
javascript
| 2018-03-26T17:48:36Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 12,411 |
["packages/react-dom/src/__tests__/ReactComponentLifeCycle-test.js", "packages/react-reconciler/src/ReactFiberClassComponent.js", "packages/react-test-renderer/src/ReactShallowRenderer.js", "packages/react-test-renderer/src/__tests__/ReactShallowRenderer-test.js", "packages/react/src/__tests__/createReactClassIntegration-test.js"]
|
UNSAFE_* Lifecycle hooks don't run if getDerivedStateFromProps is present
|
https://codesandbox.io/s/xl7k7j1k6w
Not sure if this is intended or not, the longer term existence of these hooks suggested to me that they have their purposes still, so might be needed in conjunction with gDSFP. The particular use-case that caused me to discover this was a migrating: https://github.com/react-bootstrap/react-bootstrap/blob/master/src/Dropdown.js#L137 where document focus needs to be checked _prior_ to an update flushing because the update changes visual state and drops focus. I could move this to `render()` but React has traditionally yelled at me when doing ref stuff there. Admittedly in the migrated code i don't need `findDOMNode` so it'd be fine.
|
https://github.com/facebook/react/issues/12411
|
https://github.com/facebook/react/pull/12419
|
0af384b4c33bf7d039f513e0869665824e08bb6b
|
c1308adb4bbe9e5513bbc4499e0696d97e9c897f
| 2018-03-20T20:14:42Z |
javascript
| 2018-03-22T18:16:54Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 12,389 |
["packages/react-reconciler/src/ReactFiberBeginWork.js", "packages/react-reconciler/src/__tests__/ReactNewContext-test.internal.js"]
|
propagateContextChange Stuck in an Infinite Loop
|
**Do you want to request a *feature* or report a *bug*?**
Bug
**What is the current behavior?**
In the process of converting [my Navigation router](https://grahammendick.github.io/navigation/) to the new context API I found that `propagateContextChange` sometimes gets stuck in an infinite loop. I’ve created a [JsFiddle of the problem](https://jsfiddle.net/pxzo0foq/42/). If you click the 'Name' sort Hyperlink **two times** then, on the **second click**, it enters the infinite loop. The JsFiddle is as simple as I could make it. For example, if you get rid of the 'Date of Birth' column then the problem doesn’t happen!?
**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:**
https://jsfiddle.net/pxzo0foq/42/
**What is the expected behavior?**
It should not get stuck in an infinite loop
**Which versions of React, and which browser / OS are affected by this issue? Did this work in previous versions of React?**
React version 16.3.0-alpha.2.
This is to do with the new Context API introduced in 16.3
|
https://github.com/facebook/react/issues/12389
|
https://github.com/facebook/react/pull/12402
|
e1ff342bf7f451fb995e0ea2bfb10889deef022f
|
8d0942242476024ac74d4ca2fcac9824b3124dc8
| 2018-03-16T11:25:26Z |
javascript
| 2018-03-20T13:06:59Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 12,349 |
["packages/react-dom/src/shared/DOMProperty.js"]
|
React adds "className" to DOM instead of "class" in IE11
|
<!--
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*?**
It is a bug
**What is the current behavior?**
React adds "className" to the DOM instead of "class" property.
**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:**
Open the link below in IE11 and inspect the DOM. There will be the following h1 - `<h1 className="test-ie11-bug">Hello React :)</h1>`
https://vydimitrov.github.io/react-simple-boilerplate/
Link to the test repo:
https://github.com/vydimitrov/react-simple-boilerplate
**What is the expected behavior?**
To add "class" to DOM when an element has property "className"
**Which versions of React, and which browser / OS are affected by this issue? Did this work in previous versions of React?**
React 16.3.0-alpha.1
Internet Explorer 11 - version 11.540.15063.0
It works as expected in React 16.2.0
|
https://github.com/facebook/react/issues/12349
|
https://github.com/facebook/react/pull/12353
|
fcc4f52cdded62e3b8e95d18e3e6adf17b981eed
|
280acbcb713fe0751c6230b94efb45376eb4af67
| 2018-03-09T09:16:39Z |
javascript
| 2018-03-12T17:42:17Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 12,285 |
["scripts/jest/matchers/toWarnDev.js"]
|
scripts/jest/matchers/toWarnDev: Expected and actual arguments to `jest-diff` are mixed up.
|
**Do you want to request a *feature* or report a *bug*?**
bug
**What is the current behavior?**
`toWarnDev` calls `jestDiff(a, b)` as `jestDiff(normalizedMessage, expectedMessages[0])` and `jestDiff([normalizedMessage], expectedMessages)`:
https://github.com/facebook/react/blob/94518b068bf52196abea0c83f4c9926bfe2065c6/scripts/jest/matchers/toWarnDev.js#L49-L57
[`jestDiff(a, b)`](https://github.com/facebook/jest/blob/457776b2889a9be1ce8a2c636a23417264a98d99/packages/jest-diff/src/index.js#L54) [calls](https://github.com/facebook/jest/blob/457776b2889a9be1ce8a2c636a23417264a98d99/packages/jest-diff/src/index.js#L93) [`diffStrings(a, b)`](https://github.com/facebook/jest/blob/457776b2889a9be1ce8a2c636a23417264a98d99/packages/jest-diff/src/diff_strings.js#L249-L251) where by default `a` is annotated as `'Expected'` (green), `b` as `'Received'` (red).
https://github.com/facebook/jest/blob/457776b2889a9be1ce8a2c636a23417264a98d99/packages/jest-diff/src/diff_strings.js#L86-L90
```js
const getAnnotation = (options: ?DiffOptions): string =>
chalk.green('- ' + ((options && options.aAnnotation) || 'Expected')) +
'\n' +
chalk.red('+ ' + ((options && options.bAnnotation) || 'Received')) +
'\n\n';
```
**What is the expected behavior?**
`toWarnDev` should pass the expected values into `a`, the actual (received) values into `b` of `jestDiff`:
```js
jestDiff(expectedMessages[0], normalizedMessage)
```
```js
jestDiff(expectedMessages, [normalizedMessage])
```
|
https://github.com/facebook/react/issues/12285
|
https://github.com/facebook/react/pull/12288
|
9cf3733a9a43aa53d1a3cc94b2cd015b235b84cc
|
d3e0a3aaf33dd0134a2c5296db8b6ca43cab968e
| 2018-02-25T05:29:36Z |
javascript
| 2018-06-08T12:18:22Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 12,251 |
["fixtures/dom/yarn.lock", "packages/react-dom/src/__tests__/ReactDOM-test.js", "packages/react-dom/src/__tests__/ReactDOMEventListener-test.js", "packages/react-dom/src/client/ReactDOMFiberComponent.js", "packages/react-dom/src/events/ReactBrowserEventEmitter.js"]
|
Does react still require non-toplevel submit handler?
|
**Do you want to request a *feature* or report a *bug*?**
bug? Maybe.
**What is the current behavior?**
Using non-delegated handler for submit event.
**What is the expected behavior?**
After IE9, at least I know, submit event bubbled up.
**Which versions of React, and which browser / OS are affected by this issue? Did this work in previous versions of React?**
16.2.0
|
https://github.com/facebook/react/issues/12251
|
https://github.com/facebook/react/pull/13358
|
e07a3cd28f3b360755d6832a01d724a3b7576bc4
|
725e499cfb4da60da46aa5b44c4bad29cad3d08f
| 2018-02-20T05:09:50Z |
javascript
| 2018-08-10T19:10:35Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 12,182 |
["packages/react-reconciler/src/ReactFiberNewContext.js", "packages/react-reconciler/src/__tests__/ReactNewContext-test.internal.js"]
|
unexpected ContextConsumer or ContextProvider behavior
|
**Do you want to request a *feature* or report a *bug*?**
Bug
**What is the current behavior?**
When use two or more contexts that created by `createContext`,in the child-tree of cxt2.Provider,i use cxt1.Provider to set a new context of cxt1,i expect only the child consumer got the new cxt1,but the result is that the cxt2.Provider's next sibling's child-tree got an unexpected cxt2.Provider's new context
**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:**
[createContext issue](https://codesandbox.io/s/4j3m6l2l3w)
**What is the expected behavior?**
ContextProvider only affect the children node.
**Which versions of React, and which browser / OS are affected by this issue? Did this work in previous versions of React?**
React 16.3.0-alpha.0
|
https://github.com/facebook/react/issues/12182
|
https://github.com/facebook/react/pull/12187
|
49b0ca1b836751ae78fd191286b9eb3bf2be385b
|
b5e961508709ac61f7e2de60a047238216f09a04
| 2018-02-08T06:34:51Z |
javascript
| 2018-02-08T22:23:41Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 12,177 |
["packages/react-dom/src/__tests__/refs-destruction-test.js", "packages/react-reconciler/src/ReactFiberBeginWork.js"]
|
Removing a ref callback works differently on HostComponents vs ClassComponents
|
**Do you want to request a *feature* or report a *bug*?**
Bug
**What is the current behavior?**
If you remove the ref from a mounted HostComponent on an update, it calls the original ref with null. If you remove the ref from a mounted ClassComponent on an update, it does *not* call the original ref with null.
https://codesandbox.io/s/q8k7776pz9 Note that it doesn't call ref1 (passed to FooInner as ref) with null
https://codesandbox.io/s/849538y8m8 Note that it **does** call ref1 (passed to div as ref) with null
If you *modify* the ref from either a HostComponent or a ClassComponent (change from `ref={this.ref1}` to `ref={this.ref2}`), the original ref (ref1) is called with null and the new ref (ref2) is called with the instance.
(take the sandboxes above and change the second `<Foo>` to `ref={this.ref2}`
**What is the expected behavior?**
Honestly not sure but it should be consistent. The reason this is different is because for HostComponents, [we decide to schedule a Ref effect on completeWork](https://github.com/facebook/react/blob/87ae211ccd8d61796cfdef138d1e12fb7a74f85d/packages/react-reconciler/src/ReactFiberCompleteWork.js#L464-L466) as long as `current.ref !== workInProgress.ref`. For ClassComponents, [this decision is made on beginWork ](https://github.com/facebook/react/blob/87ae211ccd8d61796cfdef138d1e12fb7a74f85d/packages/react-reconciler/src/ReactFiberBeginWork.js#L184-L188) as long as `workInProgress.ref && (!current || current.ref !== workInProgress.ref)` -- which means if workInProgress.ref is null, the behavior is different (hence this issue).
We should probably pick one and make it consistent.
**Which versions of React, and which browser / OS are affected by this issue? Did this work in previous versions of React?**
Unsure - it for sure is an issue in latest/Fiber, but I haven't dug deep enough to see what happened in <16.
|
https://github.com/facebook/react/issues/12177
|
https://github.com/facebook/react/pull/12178
|
4a20ff26ecfe9bc66941d79f7fce2c67be8ee236
|
d529d2035e3fe741acdf1f01619326588fcca697
| 2018-02-07T18:47:09Z |
javascript
| 2018-02-07T20:13:42Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 12,171 |
["packages/events/SyntheticEvent.js", "packages/react-dom/src/events/__tests__/SyntheticEvent-test.js"]
|
Remove use of Proxy for events in development
|
I think maybe we should revert https://github.com/facebook/react/pull/5947.
People already think `proxyEvent` is some kind of an API: https://github.com/facebook/react/issues/12169.
It's also annoying to view in the debugger because none of the properties show up.
Instead, we could seal the event object or something like that.
|
https://github.com/facebook/react/issues/12171
|
https://github.com/facebook/react/pull/13225
|
171e0b7d447a20b7cc90808a95478c6e5b8b6af0
|
acbb4f93f09ca37f02136f9aa6f7425fde1da2cc
| 2018-02-07T08:17:56Z |
javascript
| 2018-07-17T23:14:13Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 12,158 |
["packages/react-reconciler/src/ReactChildFiber.js", "packages/react-reconciler/src/__tests__/ReactIncrementalSideEffects-test.internal.js", "packages/react/src/__tests__/ReactStrictMode-test.internal.js"]
|
Add string-ref warning to strict mode subtrees
|
https://github.com/facebook/react/issues/12158
|
https://github.com/facebook/react/pull/12161
|
f05296baf586dc8e6d6e6a3e6d666c8dd8e4ccd5
|
6f2f55ed568063a4efbe4b2f1e11297da877fe73
| 2018-02-05T20:23:23Z |
javascript
| 2018-02-06T15:43:31Z |
|
closed
|
facebook/react
|
https://github.com/facebook/react
| 12,150 |
["packages/react-test-renderer/src/ReactTestRenderer.js", "packages/react-test-renderer/src/__tests__/ReactTestRenderer-test.js"]
|
ContextConsumer and ContextProvider in react-test-renderer
|
<!--
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*?**
bug
**What is the current behavior?**
> Unsupported component type 13 in test renderer. This is probably a bug in React.
`invariant()` is called because `react-test-renderer` does not know about the the `ContextConsumer` and `ContextProvider` component types.
**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:**
https://codesandbox.io/s/wkvl8ojpjw
**What is the expected behavior?**
It can identify the new context types.
**Which versions of React, and which browser / OS are affected by this issue? Did this work in previous versions of React?**
16.3.0-alpha.0
|
https://github.com/facebook/react/issues/12150
|
https://github.com/facebook/react/pull/12237
|
93ce76b7eafa7aaa9ebeb60efcc94c1f3954607c
|
e68c0164aa92ed383a0060639d9c449e2f740fb5
| 2018-02-04T18:59:20Z |
javascript
| 2018-02-16T19:27:20Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 12,147 |
["packages/react-art/package.json", "packages/react-call-return/package.json", "packages/react-dom/package.json", "packages/react-noop-renderer/package.json", "packages/react-reconciler/package.json", "packages/react-test-renderer/package.json", "scripts/release/build-commands/update-noop-renderer-dependencies.js", "scripts/release/build-commands/update-package-versions.js", "scripts/release/build.js"]
|
CI broken
|
Looks like commit 8a995f7 broke our CI. Not sure why. The release script passed all 4 test variants (dev+prod and bundle dev+prod). At first I was unable to reproduce the broken behavior, but after doing a fresh install (`rm -rf node_modules && yarn install`) I'm able.
I'm not sure yet what is causing this. Perhaps it has something to do with Yarn workspace and prerelease versions?
|
https://github.com/facebook/react/issues/12147
|
https://github.com/facebook/react/pull/12148
|
be85544b89cd88add5e1158812c16f4ca6c39b3a
|
dc271876a225d6efc7f13c07f82aba6cba2f39c7
| 2018-02-04T03:14:18Z |
javascript
| 2018-02-04T16:54:42Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 12,145 |
["scripts/tasks/eslint.js"]
|
process.CI check looks like a typo
|
`if (process.CI)` check looks like a typo, here
https://github.com/facebook/react/blob/master/scripts/tasks/eslint.js#L13
May be I'm wrong but I haven't found what `process.CI` is.
If CI is https://circleci.com/docs/2.0/env-vars/#circleci-environment-variable-descriptions it should be checked as `process.env.CI`
|
https://github.com/facebook/react/issues/12145
|
https://github.com/facebook/react/pull/12146
|
885a291141330eb74e8a98316e286a9ad093f22c
|
be85544b89cd88add5e1158812c16f4ca6c39b3a
| 2018-02-03T19:28:28Z |
javascript
| 2018-02-04T01:56:06Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 12,144 |
["packages/react-dom/src/__tests__/ReactComponentLifeCycle-test.internal.js", "packages/react-dom/src/__tests__/ReactDOMServerLifecycles-test.internal.js", "packages/react-dom/src/server/ReactPartialRenderer.js", "packages/react-reconciler/src/ReactStrictModeWarnings.js", "packages/react-test-renderer/src/ReactShallowRenderer.js", "packages/react-test-renderer/src/__tests__/ReactShallowRenderer-test.internal.js"]
|
Change lifecycle deprecation warnings to low-pri
|
Any warnings driven off of `warnAboutDeprecatedLifecycles` should use `shared/lowPriorityWarning` rather than `fbjs/lib/warning` to avoid thrashing automated tests.
|
https://github.com/facebook/react/issues/12144
|
https://github.com/facebook/react/pull/12159
|
86914cb30a7654d13a17d34a3b3a5af97c2d2855
|
f05296baf586dc8e6d6e6a3e6d666c8dd8e4ccd5
| 2018-02-03T17:17:38Z |
javascript
| 2018-02-05T21:39:07Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 12,106 |
["packages/react-test-renderer/src/ReactTestRenderer.js", "packages/react-test-renderer/src/__tests__/ReactTestRenderer-test.js"]
|
react-test-renderer: toTree() does not yet know how to handle nodes with tag=10
|
<!--
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*?**
bug.
**What is the current behavior?**
```js
renderer.create(
<Fragment>
<Fragment>foo</Fragment>
</Fragment>
)
.toTree()
```
> Invariant Violation: toTree() does not yet know how to handle nodes with tag=10
https://codesandbox.io/s/olky92m01y
**What is the expected behavior?**
returns a tree.
**Which versions of React, and which browser / OS are affected by this issue? Did this work in previous versions of React?**
16.2.0
|
https://github.com/facebook/react/issues/12106
|
https://github.com/facebook/react/pull/12107
|
40a9e64e1f23e478036c5f302a60e5503ccf3a30
|
ef8d6d92a28835a8ab89a876f877306a5c3feffe
| 2018-01-27T15:52:20Z |
javascript
| 2018-01-27T23:03:27Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 12,072 |
["packages/react-dom/src/__tests__/ReactDOMTextarea-test.js", "packages/react-dom/src/client/ReactDOMTextarea.js"]
|
All controlled <textarea/> fields re-render on any setState() call even though their data has not changed.
|
**Do you want to request a *feature* or report a *bug*?**
Reporting a bug.
**What is the current behavior?**
Basically I have many controlled `<textarea/>` fields in the app that I am currently developing and I normally only want them to re-render when their values have changed through their onChange event, but they re-render every time setState() gets called anywhere within the component or any parent component.
I even tried creating a custom component which only contains a `<textarea/>` and setting shouldComponentUpdate() to return false but it seems to ignore that command completely. It's weird because in this case the console.log()s in my child component's render() function don't get executed but the field gets re-rendered nonetheless.
This issue is causing some major performance problems in my app. I hope someone addresses this issue as fast as possible so I can continue working on my react project. Maybe someone can provide a temporary workaround? Please let me know if there's something I'm missing or if I'm using this field incorrectly.
**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:**
I have isolated the problem in a new empty app and you can see the behavior in the following gif:
https://gfycat.com/DentalExcitableIndri
The code is very simple. You can recreate this problem by creating 2 controlled fields, one `<input/>` and another `<textarea/>`. Then go to your browser, inspect the `<textarea/>` element and input something in the basic input field. The `<textarea/>` will re-render on every new typed letter in the `<input/>` field but not the other way around.
```javascript
import React, { Component } from 'react';
import logo from './logo.svg';
import './App.css';
class App extends Component {
constructor(props){
super(props);
this.state = {
ta1 : "ta1",
ta2 : "ta2",
ta3 : "ta3",
i1 : "i1",
i2 : "i2",
i3 : "i3"
};
this.handleInputChange = this.handleInputChange.bind(this);
}
handleInputChange(event){
this.setState({
[event.target.name] : event.target.value
});
}
render() {
return (
<div className="App">
<header className="App-header">
<img src={logo} className="App-logo" alt="logo" />
<h1 className="App-title">Welcome to React</h1>
</header>
<p className="App-intro">
test
</p>
<div className="row">
<div className="column">
textareas
<textarea value={this.state.ta1} name="ta1" onChange={this.handleInputChange}/>
<textarea value={this.state.ta2} name="ta2" onChange={this.handleInputChange}/>
<textarea value={this.state.ta3} name="ta3" onChange={this.handleInputChange}/>
</div>
<div className="column">
input fields
<input value={this.state.i1} name="i1" onChange={this.handleInputChange}/>
<input value={this.state.i2} name="i2" onChange={this.handleInputChange}/>
<input value={this.state.i3} name="i3" onChange={this.handleInputChange}/>
</div>
</div>
</div>
);
}
}
export default App;
```
**What is the expected behavior?**
The `<textarea/>` field should only re-render when its data is changed.
**Which versions of React, and which browser / OS are affected by this issue? Did this work in previous versions of React?**
I was able to recreate this problem on React versions 16.2.0 (latest) and 15.6.1. I did not test on other versions so I am not sure if this worked correctly before. I'm using Chrome 62.0.3202.94 64-bit (latest) and Firefox 57.0.4 64-bit (latest) on Ubuntu 14.04.
|
https://github.com/facebook/react/issues/12072
|
https://github.com/facebook/react/pull/13643
|
0c9c591bfb4c280d69bc9c1dd692ab0028bb0f8e
|
d92114b98e8509a9fac8ad4f56110020fa8d6987
| 2018-01-22T17:45:57Z |
javascript
| 2018-09-14T23:09:07Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 12,062 |
["fixtures/dom/src/components/fixtures/text-inputs/ReplaceEmailInput.js", "fixtures/dom/src/components/fixtures/text-inputs/index.js", "packages/react-dom/src/client/ReactInputSelection.js"]
|
Problem when input type changes from email to text
|
**Do you want to request a *feature* or report a *bug*?**
Bug
**What is the current behavior?**
When an input field changes from "email" to "text", an exception TypeError will be thrown from setSelection.
**Reproduction**
See here: https://github.com/adrianimboden/react-bug-reproduction
**What is the expected behavior?**
It should not crash because of an uncaught exception
**Which versions of React, and which browser / OS are affected by this issue? Did this work in previous versions of React?**
master (https://github.com/facebook/react/commit/4ca7855ca062d5d7dfca83c86acf46731e1e57ef)
Firefox
This would be my proposed change to fix the issue: https://github.com/adrianimboden/react/commit/db923b8c3c1cb23ad9fb150c160ce56f39dd3cae
|
https://github.com/facebook/react/issues/12062
|
https://github.com/facebook/react/pull/12135
|
79a740c6e32ca300d4e7ff55ab06de172d4237fd
|
e0a03c1b4d84453122c4e9d1a5e0cec52bef9066
| 2018-01-20T17:17:02Z |
javascript
| 2018-05-30T11:08:21Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 12,058 |
["packages/shared/__tests__/describeComponentFrame-test.js", "packages/shared/describeComponentFrame.js"]
|
__source makes component stack less useful
|
<!--
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*?**
Somewhere in between.
**What is the current behavior?**
When I enable `babel-plugin-transform-react-jsx-source` to automatically add a `__source` prop to every component in development, the component stack displayed for certain React warnings changes from displaying the (inferred) name of the rendering component to displaying the filename and line number of the occurrence.
React also has access to the file path, but [it is stripped](https://github.com/facebook/react/blob/30dac4e78de02fb427ee82013160ae875128d7a2/packages/shared/describeComponentFrame.js#L20) and only the filename is included. This seems to be based on the assumption that the name of a file always maps exactly to the name of the component it exports. In practice, many people place components in `ComponentName/index.js`, so `__source` currently makes the stack _less_ useful.
For example, compare:
```
in DownloadDropdown (created by Foo)
in Foo (created by Bar)
in div (created by Bar)
in div (created by Bar)
in div (created by Section)
in section (created by Section)
in Section (created by Bar)
in div (created by App)
in main (created by App)
...
```
to:
```
in DownloadDropdown (at index.js:53)
in Foo (at index.js:183)
in div (at index.js:182)
in div (at index.js:175)
in div (at index.js:29)
in section (at index.js:28)
in Section (at index.js:173)
in div (at index.js:26)
in main (at index.js:24)
...
```
**What is the expected behavior?**
I would like the stack to include the full file path, or at least to include the inferred component name alongside the filename.
Would you accept a PR for either option?
**Which versions of React, and which browser / OS are affected by this issue? Did this work in previous versions of React?**
React 16.2.0. This was introduced in #6771 for 15.2.0.
|
https://github.com/facebook/react/issues/12058
|
https://github.com/facebook/react/pull/12059
|
067cc24f55ef01a233be9e7564b337e35fed72ea
|
54d86eb8227255b207c8a82619d374738b542ca8
| 2018-01-20T00:32:03Z |
javascript
| 2018-08-09T02:33:30Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 12,044 |
["packages/react-dom/src/__tests__/ReactComponentLifeCycle-test.internal.js", "packages/react-reconciler/src/ReactFiberClassComponent.js", "packages/react-reconciler/src/ReactFiberScheduler.js", "packages/react-reconciler/src/ReactStrictModeWarnings.js", "packages/react/src/__tests__/createReactClassIntegration-test.internal.js"]
|
Coalesce warnings for deprecated lifecycle hooks
|
Coalesce warnings for deprecated lifecycle hooks and print single message after reconciliation (rather than one per component-type).
Relates to PR #12028 and [this comment](https://github.com/facebook/react/pull/12028#discussion_r162420575).
|
https://github.com/facebook/react/issues/12044
|
https://github.com/facebook/react/pull/12084
|
87ae211ccd8d61796cfdef138d1e12fb7a74f85d
|
6dabfca5774b9b2ceaa1faf8ad225e6dac016c31
| 2018-01-18T20:44:46Z |
javascript
| 2018-01-25T05:41:40Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 12,038 |
["packages/react-is/README.md", "packages/react-is/index.js", "packages/react-is/npm/index.js", "packages/react-is/package.json", "packages/react-is/src/ReactIs.js", "packages/react-is/src/__tests__/ReactIs-test.js", "scripts/rollup/bundles.js", "scripts/rollup/results.json", "scripts/rollup/validate/eslintrc.umd.js"]
|
Add React.isFragment api for verifying Fragment
|
<!--
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*?**
I want to request a feature.
**What is the current behavior?**
We have no API to verify a ReactNode is a React 16 Fragment. Though now we can use `React.isValidElement(instance) && typeof instance.type === 'symbol'` to distinguish it. It's verbose and seems uncertianly right.
**What is the expected behavior?**
Add api:
```
React.isFragment(object)
```
Verifies the object is a React.Fragment. Returns true or false.
**Which versions of React, and which browser / OS are affected by this issue? Did this work in previous versions of React?**
React 16.2
|
https://github.com/facebook/react/issues/12038
|
https://github.com/facebook/react/pull/12199
|
c7ce0091dcb2036e19df980247c9a4388ad75741
|
41b8c65f1e81ba48192fbb65fd5c742638ef4f82
| 2018-01-18T08:02:52Z |
javascript
| 2018-02-11T22:08:40Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 11,991 |
["packages/react-dom/src/__tests__/ReactCompositeComponent-test.js", "packages/react-dom/src/__tests__/ReactServerRendering-test.js", "packages/react-dom/src/server/ReactPartialRenderer.js", "packages/react-reconciler/src/ReactFiberBeginWork.js"]
|
Warn in server renderer if class doesn't extend React.Component but has a render() method
|
Same as https://github.com/facebook/react/issues/10103, but for the server renderer. You can look at implementation in https://github.com/facebook/react/pull/11168 and do something very similar in `ReactPartialRenderer.js`.
|
https://github.com/facebook/react/issues/11991
|
https://github.com/facebook/react/pull/11993
|
77f96ed9c389e4d07efa1da79d7d59951d37d5f1
|
e6e393b9c5221bfb1a5ddcc7221c42e96ab3baca
| 2018-01-08T15:29:56Z |
javascript
| 2018-01-09T16:24:49Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 11,975 |
["scripts/circleci/test_coverage.sh", "scripts/circleci/test_entry_point.sh"]
|
Re-enable coverage
|
See https://github.com/facebook/react/pull/11974, https://github.com/facebook/jest/issues/5239.
Jest OOM’d so I disabled it.
|
https://github.com/facebook/react/issues/11975
|
https://github.com/facebook/react/pull/11983
|
08c86dd76b8cdde2805c72a7609cc0b0c264a71e
|
26185759e4f05025b1fdbe5e238a44aefe5669d4
| 2018-01-05T19:00:01Z |
javascript
| 2018-01-08T02:27:24Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 11,964 |
["packages/react-test-renderer/src/ReactShallowRenderer.js", "packages/react-test-renderer/src/__tests__/ReactShallowRenderer-test.js"]
|
State initialization disparity between shallow renderer and ReactDOM
|
**Do you want to request a *feature* or report a *bug*?**
Bug
**What is the current behavior?**
The shallow renderer defaults a class component's state to empty object if it's not set. "Real" renderers don't seem to do this, so a component can render successfully in the shallow renderer, but fail at runtime in the browser with a null reference error.
**If the current behavior is a bug, please provide the steps to reproduce and if possible a minimal demo of the problem.**
Here is ReactDOM's behavior (crashes reading an uninitialized state value): https://codesandbox.io/s/4jv82zmp0
Here is the shallow renderer behavior (component able to render): https://codesandbox.io/s/pp733nrxzq
**What is the expected behavior?**
They initialize state in the same way.
**Which versions of React, and which browser / OS are affected by this issue? Did this work in previous versions of React?**
React 16. This wasn't an issue in React 15.
|
https://github.com/facebook/react/issues/11964
|
https://github.com/facebook/react/pull/11965
|
808f31af5cc571880058281c3ba1e1c69ce8aa5c
|
39be83565c65f9c522150e52375167568a2a1459
| 2018-01-04T22:38:24Z |
javascript
| 2018-01-05T18:44:44Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 11,963 |
["fixtures/attribute-behavior/package.json", "fixtures/dom/package.json", "fixtures/expiration/package.json", "fixtures/expiration/yarn.lock", "scripts/rollup/results.json"]
|
On running yarn start or npm start for fixtures is throwing error.
|
<!--
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*?**
A bug
**What is the current behavior?**
on running npm start or yarn start in "react/fixtures/dom is throwing errors in terminal.
**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:**
1.Go to react/fixtures/dom
2. Run yarn start or npm start
Following errors can be seen in the terminal
`> [email protected] prestart /home/gmadmin/Practice/react/fixtures/dom
> cp ../../build/dist/{react,react-dom}.development.js public/
cp: cannot stat '../../build/dist/{react,react-dom}.development.js': No such file or directory
npm ERR! code ELIFECYCLE
npm ERR! errno 1
npm ERR! [email protected] prestart: `cp ../../build/dist/{react,react-dom}.development.js public/`
npm ERR! Exit status 1
npm ERR!
npm ERR! Failed at the [email protected] prestart script.
npm ERR! This is probably not a problem with npm. There is likely additional logging output above.
npm ERR! A complete log of this run can be found in:
npm ERR! /home/gmadmin/.npm/_logs/2018-01-04T18_32_47_442Z-debug.log
`
Note: I have already run npm run build in the react parent directory and i can see react.development.js and react-dom.development.js in react/build/dist/. Running "cp ../../build/dist/{react,react-dom}.development.js public/" this command on terminal works fine. But when it fails when run through "prestart" command in package.json.
**What is the expected behavior?**
copying should happen successfully and npm start should successfully turn up the server.
**Which versions of React, and which browser / OS are affected by this issue?
React "version": "16.2.0",
kernel : "4.4.0-53-generic GNU/Linux"
os info: {
Distributor ID: LinuxMint
Release: 18.1
Codename: serena}
Did this work in previous versions of React?
Don't know
If its a bug, let me work on it.
|
https://github.com/facebook/react/issues/11963
|
https://github.com/facebook/react/pull/11982
|
13c5e2b53112c1cd77dd52e028a4bb33dc492d1b
|
77f96ed9c389e4d07efa1da79d7d59951d37d5f1
| 2018-01-04T19:09:51Z |
javascript
| 2018-01-09T11:15:09Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 11,955 |
["CHANGELOG.md"]
|
`react-call-return`: Expected to find a host parent.
|
<!--
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*?**
Bug
**What is the current behavior?**
A specific order of unmounting and remounting `unstable_createReturn`s from `react-call-return` causes an invariant violation in `unmountHostComponents`.
**Reproduce**
The following sandbox example crashes with an invariant violation when both the `min` and `cycle` props are *odd* numbers greater than zero.
https://codesandbox.io/s/llyjz19rz7
**What is the expected behavior?**
The app does not crash and cycles the number of items in the list.
**Which versions of React, and which browser / OS are affected by this issue? Did this work in previous versions of React?**
`react` and `react-dom` versions 16.1 and newer, `react-call-return` version 0.5.0
|
https://github.com/facebook/react/issues/11955
|
https://github.com/facebook/react/pull/12488
|
8e3d94ffa1d2e19a5bf4b9f8030973b65b0fc854
|
2c3f5fb97b6ea077f3e9aae6c6587bfe8328036d
| 2018-01-03T15:16:26Z |
javascript
| 2018-03-29T20:56:45Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 11,918 |
["packages/react-dom/src/client/ReactDOMFiberComponent.js", "packages/react-dom/src/client/ReactDOMHostConfig.js"]
|
Click events don't bubble from Portal content on mobile Safari
|
**Do you want to request a *feature* or report a *bug*?**
Bug
**What is the current behavior?**
In mobile Safari, DOM events originating from elements rendered through a Portal do **not** bubble (propagate) up the React component tree.
**If the current behavior is a bug, please provide the steps to reproduce....**
Please see the following JSFiddle: https://jsfiddle.net/zebulonj/wf7orwan/
In this JSFiddle, the expected behavior (seen in a desktop browser) is that clicking on the portal content (appearing as a modal, including backdrop) should toggle to the "On" or "Off" indicator rendered in the main content. However, in mobile Safari (iOS 11.1.1), the click events do not propagate.
**What is the expected behavior?**
Click events should propagate up the React component tree.
**Which versions of React, and which browser / OS are affected by this issue? Did this work in previous versions of React?**
Affected: React 16.2
iOS 11.1.1
|
https://github.com/facebook/react/issues/11918
|
https://github.com/facebook/react/pull/11927
|
0da5102cf02890a4e006f4d949ae4fbf2bfd85af
|
b3a4cfea577e3333941a7cb2b01d2b957988e00b
| 2017-12-24T17:20:21Z |
javascript
| 2018-08-18T01:10:20Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 11,911 |
["fixtures/dom/src/components/fixtures/selects/index.js", "packages/react-dom/src/__tests__/ReactDOMOption-test.js", "packages/react-dom/src/__tests__/ReactDOMSelect-test.js", "packages/react-dom/src/__tests__/ReactDOMServerIntegrationForms-test.js", "packages/react-dom/src/__tests__/utils/ReactDOMServerIntegrationTestUtils.js", "packages/react-dom/src/client/ReactDOMFiberOption.js", "packages/react-dom/src/client/ReactDOMHostConfig.js"]
|
React DOM crashes when <option> contains three interpolated value if one is a conditional.
|
<!--
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*?**
Bug
**What is the current behavior?**
React DOM crashes when `<option>` contains three interpolated value if one is a conditional.
Reproduction: https://jsfiddle.net/0opjvycp/
1. Change the value of the `<select>`
2. React crashes with `NotFoundError: Node was not found`
From what I can tell, the conditional value is necessary, and it must be three interpolated values.
**What is the expected behavior?**
React should not crash.
**Which versions of React, and which browser / OS are affected by this issue? Did this work in previous versions of React?**
React DOM 16.2 and 16.0. I think this worked in 15.6 - https://jsfiddle.net/mrwkmuqc/ does not crash
|
https://github.com/facebook/react/issues/11911
|
https://github.com/facebook/react/pull/13261
|
2a2ef7e0fd86612404676b15bb3dd9df7ade536e
|
0182a74632b66a226921ca13dcfe359c67e15c6f
| 2017-12-22T12:10:29Z |
javascript
| 2018-08-01T15:16:34Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 11,902 |
["packages/react-dom/src/__tests__/ReactDOMServerIntegrationElements-test.js", "packages/react-dom/src/server/ReactPartialRenderer.js"]
|
Provide a better message when component type is undefined with SSR
|
This might be a reference to this one: https://github.com/facebook/react/issues/10212#issuecomment-346302374
The problem still exist.
So, trying to render with server side and got this error. I don't have this notation in my code but this might be a problem with dependencies.
```
TypeError: Cannot read property 'toLowerCase' of undefined
at ReactDOMServerRenderer.renderDOM (/Users/lukasborawski/Dev/web-client/node_modules/react-dom/cjs/react-dom-server.node.development.js:2304:27)
...
```
```
"react": "^16.1.1",
"react-dom": "^16.2.0",
```
Thx for help.
|
https://github.com/facebook/react/issues/11902
|
https://github.com/facebook/react/pull/11966
|
f828ca407febc6287011f8ae89aa469c4d522fcc
|
86914cb30a7654d13a17d34a3b3a5af97c2d2855
| 2017-12-21T14:55:26Z |
javascript
| 2018-02-05T17:09:09Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 11,862 |
["packages/react-test-renderer/src/ReactShallowRenderer.js", "packages/react-test-renderer/src/__tests__/ReactShallowRenderer-test.js"]
|
ShallowRenderer not considering contextTypes when rendering children
|
**Do you want to request a *feature* or report a *bug*?**
Bug
**What is the current behavior?**
ShallowRenderer not considering contextTypes.
**Note: I was not able to easily include react-test-renderer in a fiddle so I include a unit test.
```
//Context used in shallow and normal rendering
const context = { appName: 'abc', mode: 'def' };
class TopLevelComponent extends React.Component {
render() { return <div>{this.props.children}</div>; }
getChildContext() { return context; }
}
TopLevelComponent.childContextTypes = { appName: PropTypes.string, mode: PropTypes.string };
class Component1 extends React.Component {
render() { return <div mode={this.context.mode} appname={this.context.appName}/>; }
}
Component1.contextTypes = { appName: PropTypes.string };
//Use shallow renderer and assert that we got appName but not mode
const shallowComponent1 = new ShallowRenderer().render(<Component1/>, context);
expect(shallowComponent1.props.appname).toBe('abc'); //Pass
expect(shallowComponent1.props.mode).toBeUndefined(); //This expectation fails
//Render Component1 under a component with childContext
const wrappedEl = React.createElement(() => <TopLevelComponent><Component1/></TopLevelComponent>);
const domNode = document.createElement('div');
ReactDOM.render(wrappedEl, domNode);
const html = domNode.innerHTML;
expect(html).toContain('abc'); //Pass
expect(html).not.toContain('def'); //Pass: As you can see this property is undefined
```
**What is the expected behavior?**
I would expect that when rendering Component1 with ShallowRenderer, render method would check Component1.contextTypes and mask any property that is not specified there just like function getMaskedContext(workInProgress, unmaskedContext) does in react-dom library.
I fixed this behavior by changing render method like this: (inpired again by getMaskedContext)
```
const oldRender = ShallowRenderer.prototype.render;
ShallowRenderer.prototype.render = function render(element) {
const unmaskedContext = arguments.length > 1 && arguments[1] !== undefined ? arguments[1] : {};
if (element && element.type && unmaskedContext) {
if (element.type.hasOwnProperty('contextTypes')) {
const context = {};
for (const key in element.type.contextTypes) {
context[key] = unmaskedContext[key];
}
arguments[1] = context;
}
}
return oldRender.apply(this, arguments);
};
```
**Which versions of React, and which browser / OS are affected by this issue? Did this work in previous versions of React?**
This bug is for version 16. We are upgrading from React 15, we were using enzyme shallow rendering and this wasn't the behavior. It is difficult to say which part of the code has changed between React 15 and enzyme at this point but my code sample includes only code from react libraries.
|
https://github.com/facebook/react/issues/11862
|
https://github.com/facebook/react/pull/11922
|
39be83565c65f9c522150e52375167568a2a1459
|
9d310e0bc7b9d5ce39d82536dfcb67f98462a346
| 2017-12-15T12:26:55Z |
javascript
| 2018-01-05T18:49:57Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 11,858 |
["fixtures/attribute-behavior/AttributeTableSnapshot.md", "fixtures/attribute-behavior/src/attributes.js", "packages/react-dom/src/shared/DOMProperty.js", "packages/react-dom/src/shared/possibleStandardNames.js"]
|
add support for nomodule attribute on the script tag
|
<!--
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*?**
feature: add support for the [nomodule](https://hospodarets.com/native-ecmascript-modules-nomodule) attribute to the script tag.
**What is the current behavior?**
with `<script nomodule src="foo.js" />` i get:
```
Warning: Received `true` for a non-boolean attribute `nomodule`.
If you want to write it to the DOM, pass a string instead: nomodule="true" or nomodule={value.toString()}.
```
and the attribute is removed.
**If the current behavior is a bug, please provide the steps to reproduce and if possible a minimal demo of the problem via https://jsfiddle.net or similar (template for React 16: https://jsfiddle.net/Luktwrdm/, template for React 15: https://jsfiddle.net/hmbg7e9w/).**
https://jsfiddle.net/hzdkwne6/
**What is the expected behavior?**
the `nomodule` attribute should be respected
**Which versions of React, and which browser / OS are affected by this issue? Did this work in previous versions of React?**
react 16.2.0
chrome
osx
not a regression, i don't think it's ever worked before.
if this isn't an intentional oversight i'll open a pr to add it, seems like it's just a change to https://github.com/facebook/react/blob/master/fixtures/attribute-behavior/src/attributes.js ?
|
https://github.com/facebook/react/issues/11858
|
https://github.com/facebook/react/pull/11900
|
faa4218632ca21f3af731fb6c14849fc054fca15
|
9ff3ce67eaf07e8f62704133a8d748f1c0679a59
| 2017-12-14T19:52:27Z |
javascript
| 2017-12-22T18:18:51Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 11,807 |
["packages/react-dom/src/__tests__/ReactServerRenderingHydration.js", "packages/react-dom/src/client/ReactDOMComponent.js"]
|
False positive warning about style mismatch when hydrating server markup in IE11
|
<!--
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*?**
Bug? (not a biggy - not sure if IE11 is support for development)
**What is the current behavior?**
IE11 seems to always throw a warning for hydration failures
eg
> **Warning: Prop `style` did not match. Server: "text-decoration: none;" Client: "text-decoration:none"**
Note: there is a space between the colon and none from Server, Client has no space
**If the current behavior is a bug, please provide the steps to reproduce and if possible a minimal demo of the problem via https://jsfiddle.net or similar (template for React 16: https://jsfiddle.net/Luktwrdm/, template for React 15: https://jsfiddle.net/hmbg7e9w/).**
Have a SSR rendered component, using inline styles
```
<a href="/page1" style={{ textDecoration: "none" }}>
<h1 className="splash-head">Go to Page1</h1>
</a>
```
Then hydrate it on IE11
`ReactDOM.hydrate(component, document.getElementById("app"));`
This **doesnt** occur in either Chrome or Firefox
**What is the expected behavior?**
No warning
**Which versions of React, and which browser / OS are affected by this issue? Did this work in previous versions of React?**
React: 16.1.1
Browser: IE11
|
https://github.com/facebook/react/issues/11807
|
https://github.com/facebook/react/pull/13534
|
25d48a7281bcb2c51927c562a43295b8ce0bd983
|
2d4705e7532d32147da8ee61e5760124d068a8a6
| 2017-12-08T02:20:24Z |
javascript
| 2018-09-04T13:26:51Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 11,795 |
["packages/react-dom/src/__tests__/ReactDOMSelect-test.js", "packages/react-dom/src/client/ReactDOMFiberOption.js"]
|
Select warning fires repeatedly
|
<!--
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*?**
Bug
**What is the current behavior?**
See fiddle https://jsfiddle.net/j2nzg31L/1/
The warning related to the select fires repeatedly. @gaearon tells me it should dedupe, and that this is a bug.
**If the current behavior is a bug, please provide the steps to reproduce and if possible a minimal demo of the problem via https://jsfiddle.net or similar (template for React 16: https://jsfiddle.net/Luktwrdm/, template for React 15: https://jsfiddle.net/hmbg7e9w/).**
See above
**What is the expected behavior?**
See above
**Which versions of React, and which browser / OS are affected by this issue? Did this work in previous versions of React?**
16.2
|
https://github.com/facebook/react/issues/11795
|
https://github.com/facebook/react/pull/11821
|
f23dd7150f270c843707f1a06ac12a6af63b7320
|
51e3f498a20572ecd5360c0460da212903a089bb
| 2017-12-07T16:52:54Z |
javascript
| 2017-12-10T02:06:41Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 11,789 |
["packages/react-dom/src/__tests__/ReactDOMServerIntegrationElements-test.js", "packages/react-dom/src/__tests__/ReactDOMServerIntegrationReconnecting-test.js", "packages/react-dom/src/client/ReactDOMFiberComponent.js", "packages/react-dom/src/client/ReactDOMHostConfig.js"]
|
hydrating a component with `dangerouslySetInnerHTML` and `toString` causes a warning, and the component to not render
|
**Do you want to request a *feature* or report a *bug*?**
bug
**What is the current behavior?**
When hydrating a pre-rendered component that uses `dangerouslySetInnerHTML`, + an object with a `toString` method, we get a warning such as:
```
Warning: Did not expect server HTML to contain the text node "Bonjour" in <p>.
```
The component will also appear blank.
This is noteworthy since:
- The markup returned by `render` appears to be correct, meaning:
- The markup that an SSR server returns is correct
- The markup that appears on page if we just use `render` on the client, without SSR is correct
- It's only if we do both, that we get errors.
**If the current behavior is a bug, please provide the steps to reproduce and if possible a minimal demo of the problem via https://jsfiddle.net or similar (template for React 16: https://jsfiddle.net/Luktwrdm/, template for React 15: https://jsfiddle.net/hmbg7e9w/).**
https://codesandbox.io/s/vqkq34o965
(Code, for posterity:)
```js
import React from "react";
import { hydrate } from "react-dom";
const HelloObj = {
toString: () => "Bonjour"
};
const MyComponent = () => (
<p
dangerouslySetInnerHTML={{
__html: HelloObj
}}
/>
);
hydrate(<MyComponent />, document.getElementById("app"));
```
**What is the expected behavior?**
The component is not blank, and no error is outputted in the console
**Which versions of React, and which browser / OS are affected by this issue? Did this work in previous versions of React?**
- react/react-dom 16.2.0
- This worked < 16 (confirmed working in 15.6.1, if we use the `render` method)
- Chrome Version 62.0.3202.94
- OS X 10.12.6
|
https://github.com/facebook/react/issues/11789
|
https://github.com/facebook/react/pull/13353
|
0072b59984f0743b95664557c17904e8c0912ec5
|
be4533af7d29f020499d3c01931e55ef8666a419
| 2017-12-07T00:21:56Z |
javascript
| 2018-08-09T17:05:05Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 11,759 |
["packages/react/src/ReactElementValidator.js"]
|
"React.createElement: type is invalid" warning in IE11 when using React.Fragment
|
https://codesandbox.io/embed/j30o67kpxy
React shows warning in console if I use React.Fragment
```
Warning: React.createElement: type is invalid -- expected a string (for built-in components) or a class/function (for composite components) but got: object.
```
|
https://github.com/facebook/react/issues/11759
|
https://github.com/facebook/react/pull/11823
|
abdbb16d4bec5c444c2d8e69b508871ab1335793
|
a5025b161048c4e7211cf4c7298df7fcfc03bfa0
| 2017-12-04T08:56:34Z |
javascript
| 2017-12-11T03:25:28Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 11,726 |
["packages/react-dom/src/__tests__/ReactServerRendering-test.js", "packages/react-dom/src/client/ReactDOM.js"]
|
client/server mismatch for text inside of an a/button causes the the a/button to focus on page load
|
**Do you want to request a *feature* or report a *bug*?**
bug
**What is the current behavior?**
If there is a client/server mismatch for text inside of an a or button element, it causes it to be automatically focused on page load.
**If the current behavior is a bug, please provide the steps to reproduce and if possible a minimal demo of the problem via https://jsfiddle.net or similar (template for React 16: https://jsfiddle.net/Luktwrdm/, template for React 15: https://jsfiddle.net/hmbg7e9w/).**
```
let text;
let URL = '/randomURL';
if (typeof window !== 'undefined') {
text = 'CLIENT TEXT';
} else {
text = 'TEXT';
}
return (
<a href={URL}>
{text}
</a>
);
```
if the focus is not on the window (i.e. if a dom element is selected in the chrome inspector or react inspector widnow), then it doesn't add the focus.
**What is the expected behavior?**
The expected behavior is that it would either show CLIENT TEXT or TEXT (depending on how React16 resolves the diff), but for it not to be automatically highlighted
**Which versions of React, and which browser / OS are affected by this issue? Did this work in previous versions of React?**
React16 / macosx / chrome + ff + safari
|
https://github.com/facebook/react/issues/11726
|
https://github.com/facebook/react/pull/11737
|
5bd2321ae3dd7d68ac02dee3c3f271e9d0ee8784
|
19bc2dd090c44be862bae59c64e89acb8d0827f5
| 2017-11-30T19:53:33Z |
javascript
| 2017-12-06T15:23:37Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 11,692 |
["packages/react-dom/src/__tests__/ReactServerRendering-test.js", "packages/react-dom/src/server/ReactPartialRenderer.js"]
|
ReactDOMServer.renderToStaticMarkup gets stuck while trying to render Portals
|
Calling `ReactDOMServer.renderToStaticMarkup(elementWithPortal);` gets stuck in a loop while trying to do a static render.
Current behavior:
The behavior can be seen here: https://codepen.io/anon/pen/BmqWOM by un-commenting line 58 (warning: this will make your tab freeze)
Expected behavior:
It should return with results or throw some kind of error if the input is not valid
|
https://github.com/facebook/react/issues/11692
|
https://github.com/facebook/react/pull/11709
|
e0c31137434d7c7df2bc3475cb7682b1d7eecc8b
|
c3f1b6cd91ffa3e05465c3ed2f48225876c1bd3b
| 2017-11-28T21:36:13Z |
javascript
| 2017-11-29T21:45:38Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 11,687 |
["fixtures/dom/src/components/fixtures/error-handling/index.js", "packages/shared/invokeGuardedCallback.js"]
|
react-dom: Ability to access window.event in development
|
Imported from https://github.com/facebook/react/issues/10474#issuecomment-345331144 at request of @gaearon
**Do you want to request a *feature* or report a *bug*?** Bug
**What is the current behavior?**
In development mode React uses `invokeGuardedCallback` to be able to catch errors without interrupting the normal "break on uncaught exceptions" feature of devtools.
In the current implementation of `invokeGuardedCallback` an event is created and dispatched, which allows for error handling without a `try{ } catch` block.
Unfortunately it has the side effect of overriding `window.event`, and so you are unable to access `window.event` within guarded callbacks when running React in development mode. In production mode it works fine.
We would like to access window.event at Superhuman to be able to detect where DOM focus is from the user's point of view. Usually this is document.activeElement, but in a few cases (like during a blur event, or when clicking between two iframes) the focus will end up in a different place.
This works fine in production, and we currently work around this in development mode by overriding `ReactDOM.__SECRET_INTERNALS_DO_NOT_USE_OR_YOU_WILL_BE_FIRED.ReactDOMEventListener.handleTopLevel` and maintiaining a reference to the event. In production mode we just use window.event.
**If the current behavior is a bug, please provide the steps to reproduce and if possible a minimal demo of the problem via https://jsfiddle.net or similar (template for React 16: https://jsfiddle.net/Luktwrdm/, template for React 15: https://jsfiddle.net/hmbg7e9w/).**
in development: https://jsfiddle.net/n782L2qg/, in production: https://jsfiddle.net/kqt2o7pr/. In both cases I would expect the alert to say `window.event.type: click`, but in development it says `window.event.type: react-invokeguardedcallback`
**What is the expected behavior?**
I would like a way to access window.event in development. I'm happy for this to be somewhat obscure, but ideally I would not have to use the secret internals to make this work.
**Which versions of React, and which browser / OS are affected by this issue? Did this work in previous versions of React?** Development 0.15 and 0.16 at least. probably before.
|
https://github.com/facebook/react/issues/11687
|
https://github.com/facebook/react/pull/11696
|
ade4dd3f6fb542677627001fe7e3477b27de96e2
|
69e2a0d732e1ca74f6dc5df9d0ddd0bf24373965
| 2017-11-28T20:33:31Z |
javascript
| 2018-08-14T20:35:31Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 11,660 |
["package.json", "scripts/circleci/build.sh", "scripts/release/build-commands/run-automated-bundle-tests.js", "scripts/release/build-commands/run-automated-tests.js", "scripts/release/build.js", "scripts/release/utils.js"]
|
Add bundle linting and tests to the release script
|
On CI, we currently lint bundles:
https://github.com/facebook/react/blob/d1cb28c86b966cb33c373b532abe62dfe563d00e/scripts/circleci/build.sh#L11-L12
and then run both development and production bundle tests:
https://github.com/facebook/react/blob/d1cb28c86b966cb33c373b532abe62dfe563d00e/scripts/circleci/test_entry_point.sh#L24-L25
We should add the first step (linting bundles) into a `package.json` script called `lint-build`. Then we can change the CI command to run `yarn lint-build`.
Finally, we should add both `yarn lint-build`, `yarn test-prod`, `yarn test-build`, and `yarn test-build-prod` to the release script [into this function](https://github.com/facebook/react/blob/d1cb28c86b966cb33c373b532abe62dfe563d00e/scripts/release/build-commands/run-automated-tests.js#L23).
<s>Note that this function is already broken. We need to replace `runYarnTask(cwd, 'jest', 'Jest failed'),` with `runYarnTask(cwd, 'test', 'Jest failed'),` too since just running `jest` no longer works due to a custom config. Just running `test` should work because we have a custom `test` script.</s> Never mind, I fixed this in https://github.com/facebook/react/commit/f53bd033e79994dbda112a091ddbf784a3c30aaf.
To verify your changes, you can run `./scripts/release/build.js -v 16.2.0` and see that the whole workflow can run successfully.
This task requires understanding of how `scripts` in `package.json` work.
|
https://github.com/facebook/react/issues/11660
|
https://github.com/facebook/react/pull/11662
|
f53bd033e79994dbda112a091ddbf784a3c30aaf
|
53ef71b8e8ba03058f99d57fcb1efe15956c3f36
| 2017-11-25T16:08:00Z |
javascript
| 2017-11-26T16:47:20Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 11,657 |
["circle.yml", "scripts/circleci/test_entry_point.sh"]
|
Investigate CI failure (possible race condition)
|
It looks like this:
```
./scripts/circleci/upload_build.sh
curl: (26) couldn't open file "build/dist/react.development.js"
./scripts/circleci/upload_build.sh returned exit code 26
Action failed: ./scripts/circleci/upload_build.sh
```
https://circleci.com/gh/facebook/react/7816?utm_campaign=vcs-integration-link&utm_medium=referral&utm_source=github-build-link
It happened two times on random commits, and I don't see a pattern. Probably caused by me changing how tasks are distributed between parallel nodes in https://github.com/facebook/react/pull/11633. I don't understand why `upload_build.sh` would try to run before a build exists though (if that's what's happening?)
|
https://github.com/facebook/react/issues/11657
|
https://github.com/facebook/react/pull/11666
|
53ef71b8e8ba03058f99d57fcb1efe15956c3f36
|
0c164bb4851e78e5f789dd8619f17ffcfee0221f
| 2017-11-25T03:17:17Z |
javascript
| 2017-11-26T18:03:30Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 11,656 |
["packages/react-dom/src/__tests__/InvalidEventListeners-test.js", "packages/react-dom/src/__tests__/ReactTestUtils-test.js", "packages/react-dom/src/test-utils/ReactTestUtils.js"]
|
Deprecate TestUtils.SimulateNative
|
I'm not sure if it was ever intended to be a part of public API.
It's not documented, and its naming is misleading because it actually doesn't simulate a native event (oops!). It only simulates what React would do if it *received* that native event. It does that in a way that's very tightly coupled to the event system internals which is unfortunate. It is much better if we can educate people how to dispatch actual browser events in tests instead.
I think we should just deprecate it. We've removed most tests relying on it, and can leave a test or two until it can be deleted in the next major.
|
https://github.com/facebook/react/issues/11656
|
https://github.com/facebook/react/pull/13407
|
4123d729e53d11f8a29c7665ce0f52308cca3bd4
|
e55855e7aee04b8e47edc319fc828b6cff43beed
| 2017-11-25T02:56:16Z |
javascript
| 2020-04-02T22:48:03Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 11,618 |
["packages/react-dom/src/__tests__/ReactDOMServerIntegration-test.js", "packages/react-dom/src/server/ReactPartialRenderer.js", "packages/react-reconciler/src/ReactFiberClassComponent.js", "packages/react/src/__tests__/ReactCoffeeScriptClass-test.coffee", "packages/react/src/__tests__/ReactES6Class-test.js", "packages/react/src/__tests__/ReactTypeScriptClass-test.ts"]
|
Investigate uncovered differences between development and production
|
Follow up from https://github.com/facebook/react/pull/11616.
Here's one I found. These two invariants:
https://github.com/facebook/react/blob/7e7127387ba64f57299eaff8656542cea6fb3a2d/packages/react-reconciler/src/ReactFiberClassComponent.js#L330-L343
are in a function that's behind a DEV block:
https://github.com/facebook/react/blob/7e7127387ba64f57299eaff8656542cea6fb3a2d/packages/react-reconciler/src/ReactFiberClassComponent.js#L447-L449
I don't think this is intentional, but need to verify.
There may be more. I'll use this issue to track while https://github.com/facebook/react/pull/11616 preserves existing behavior.
|
https://github.com/facebook/react/issues/11618
|
https://github.com/facebook/react/pull/11630
|
1cb6199d22af6f6ba2f55e4db18ed2f4216aaaf2
|
913a125ad51777ffec91eedefa03c534a7590395
| 2017-11-22T02:31:49Z |
javascript
| 2017-11-22T18:58:53Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 11,615 |
["scripts/tasks/linc.js"]
|
yarn linc shouldn't fail because of file-has-not-changed warning
|
Just saw this after modifying ESLint config and running `yarn linc`:
```
/Users/gaearon/p/react/.eslintrc.js
0:0 warning File ignored by default. Use a negated ignore pattern (like "--ignore-pattern '!<relative/path/to/filename>'") to override
✖ 1 problem (0 errors, 1 warning)
Lint failed for changed files.
error Command failed with exit code 1.
info Visit https://yarnpkg.com/en/docs/cli/run for documentation about this command.
react)
```
We should not treat this particular warning as an error.
|
https://github.com/facebook/react/issues/11615
|
https://github.com/facebook/react/pull/11641
|
b542f42a0ff658eb383ac3b058c9805ade0ef034
|
7788bcdb26844a19c23d0f6d5ab4d20822e59c85
| 2017-11-21T20:40:42Z |
javascript
| 2017-11-28T13:54:46Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 11,593 |
["packages/react-dom/src/__tests__/ReactComponentLifeCycle-test.js", "packages/react-reconciler/src/ReactFiberClassComponent.js"]
|
Warn if this.state is set to this.props referentially
|
See this example: https://github.com/facebook/react/issues/11508#issuecomment-345464797.
I think it probably reflects a misunderstanding of how props and state work, and we should detect and warn if we see this.
|
https://github.com/facebook/react/issues/11593
|
https://github.com/facebook/react/pull/11658
|
29287f08867e72c0f4a6a6f7523e966009520308
|
672e859d311536ddf5f09fb47dcb84007703aa65
| 2017-11-18T19:24:43Z |
javascript
| 2018-08-28T13:17:44Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 11,544 |
["scripts/release/utils.js"]
|
Ignore system directories in the build script
|
Just noticed that running `./scripts/release/build.js -v 16.1.1` gives me:
```
ERROR ENOTDIR: not a directory, open '/Users/gaearon/p/react/packages/.DS_Store/package.json'
Error:
at Object.fs.openSync (fs.js:653:18)
at fs.readFileSync (fs.js:554:33)
at readdirSync.filter.dir (/Users/gaearon/p/react/scripts/release/utils.js:31:36)
at Array.filter (<anonymous>)
at getPublicPackages (/Users/gaearon/p/react/scripts/release/utils.js:29:36)
```
We should ignore system directories there.
|
https://github.com/facebook/react/issues/11544
|
https://github.com/facebook/react/pull/11546
|
200db83850257fe2e18d0212a1f96c59c02deb99
|
2d23a4563ef2bec7d90c4a7edff2657c890b4334
| 2017-11-13T15:58:01Z |
javascript
| 2017-11-13T18:21:17Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 11,526 |
["packages/react-dom/server.browser.js", "packages/react-dom/server.node.js", "packages/react-dom/src/server/ReactDOMServerBrowser.js", "packages/react-dom/src/server/ReactDOMServerNode.js"]
|
Transpilation in v16.1.0 freezes the react-dom/server interface
|
As initially reported in https://github.com/facebook/react/issues/11520#issuecomment-343593640, attempting to upgrade from v16.0 to v16.1 caused a test that was previously working to fail. The test was using [sinon](http://sinonjs.org/) to stub `renderToStaticMarkup` on `react-dom/server` and v16.1 results in the following error:
```
TypeError: Cannot redefine property: renderToStaticMarkup
at Function.defineProperty (<anonymous>)
at wrapMethod (node_modules/sinon/lib/sinon/util/core/wrap-method.js:92:16)
at stub (node_modules/sinon/lib/sinon/stub.js:60:44)
at Object.stub (node_modules/sinon/lib/sinon/collection.js:102:33)
at Context.setup (test/unit/....)
```
the export from `react-dom/cjs/react-dom-server.node.development` is frozen as:
```js
Object.freeze({
renderToString: renderToString,
renderToStaticMarkup: renderToStaticMarkup,
renderToNodeStream: renderToNodeStream,
renderToStaticNodeStream: renderToStaticNodeStream,
version: version
});
```
[the source](https://github.com/facebook/react/blob/13c491a8282cb102c40d4ed74738ebebe7901a5f/packages/react-dom/src/server/ReactDOMServerNode.js#L17-L22) is not frozen, and https://github.com/facebook/react/issues/11520#issuecomment-343604002 further confirms that the `Object.freeze()` is a result of the transpilation.
for now, i am simply splitting this report into a separate issue from https://github.com/facebook/react/issues/11520, but i will add a reproduction soon and plan to submit a PR as a follow up.
|
https://github.com/facebook/react/issues/11526
|
https://github.com/facebook/react/pull/11531
|
2fe3494f0d6d42f8a0a4dd7a196d01a191b93f90
|
901a091fe04ca5ad1fb9f825e9df39881a87debc
| 2017-11-11T04:00:34Z |
javascript
| 2017-11-13T13:52:59Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 11,519 |
["packages/shared/__tests__/describeComponentFrame-test.js", "packages/shared/describeComponentFrame.js"]
|
Consider including directory name into the stack trace for some files
|
See https://twitter.com/deadcoder0904/status/929050109425229824
Maybe if the file is called `index.js` (pretty common with some directory structures) we should include the directory name into the short component stack trace.
|
https://github.com/facebook/react/issues/11519
|
https://github.com/facebook/react/pull/12059
|
067cc24f55ef01a233be9e7564b337e35fed72ea
|
54d86eb8227255b207c8a82619d374738b542ca8
| 2017-11-10T19:51:16Z |
javascript
| 2018-08-09T02:33:30Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 11,496 |
["packages/react-dom/src/__tests__/ReactCompositeComponent-test.js", "packages/react-test-renderer/src/ReactShallowRenderer.js", "packages/react-test-renderer/src/__tests__/ReactShallowRenderer-test.js"]
|
SetState callback called before component state is updated in React Shallow Renderer
|
**Do you want to request a *feature* or report a *bug*?**
report a bug
**What is the current behavior?**
When I call setState with a callback in a test using react shallow renderer (via enzyme), the callback gets called and `this.state` is still the old state.
EDIT: This seems limited to componentWillMount
**If the current behavior is a bug, please provide the steps to reproduce and if possible a minimal demo of the problem via https://jsfiddle.net or similar (template for React 16: https://jsfiddle.net/Luktwrdm/, template for React 15: https://jsfiddle.net/hmbg7e9w/).**
EDIT: Reproduce with this https://github.com/bdwain/setstate-callback-bug
**What is the expected behavior?**
When the setState callback gets called, it should have access to the new state.
**Which versions of React, and which browser / OS are affected by this issue? Did this work in previous versions of React?**
Version 16.
It worked in v15 with enzyme 2.
EDIT: I think this is because in componentWillMount, [this line](https://github.com/facebook/react/blob/master/packages/react-test-renderer/src/ReactShallowRenderer.js#L67) causes the render method in the shallow wrapper to return early, before it updates the state.
|
https://github.com/facebook/react/issues/11496
|
https://github.com/facebook/react/pull/11507
|
913a125ad51777ffec91eedefa03c534a7590395
|
f114bad09f7a3ddc11265741cb84250fd08d5890
| 2017-11-09T00:35:37Z |
javascript
| 2017-11-22T22:15:11Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 11,481 |
["scripts/release/publish-commands/print-post-publish-summary.js", "scripts/release/publish-commands/publish-to-npm.js"]
|
Release script follow-up
|
- [X] #11482: Add a timeout before querying npm right after publish
- [X] #11482: Conditionally log some post publish steps
- [X] #11482: Maybe print ready-to-paste `yarn add ...` instructions for CRA prerelease testing
- [x] #11504: Publish all packages that aren't `private` and remove the whitelist in config
- [x] #11504: Add "v" prefix for release tags
- [x] #11504: Nit: seconds => minutes for long tasks 😉
|
https://github.com/facebook/react/issues/11481
|
https://github.com/facebook/react/pull/11482
|
a2ec771360966f7a9b91698a7626931cbc7f9105
|
8a0285fb43a3109aa7cd7b31ad91b007910fc5f8
| 2017-11-07T15:06:27Z |
javascript
| 2017-11-08T19:59:26Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 11,460 |
["packages/react-dom/src/shared/DOMProperty.js"]
|
Check .length before accessing characters
|
See comment here: https://github.com/facebook/react/pull/11457#discussion_r148954193.
Although it seems very unlikely that someone would pass a one-character prop, the price (at least in versions of V8 in Chrome right now) is reported to be about 10x.
So we should check `.length` first and then do the comparison.
**Please comment if you want to take this issue.**
**If somebody already commented, please let them finish it.**
|
https://github.com/facebook/react/issues/11460
|
https://github.com/facebook/react/pull/11462
|
92b7b172cce9958b846844f0b46fd7bbd8c5140d
|
e2e7fcce7e137e6e6fbc23cfadea1a0abaa0e6ce
| 2017-11-05T12:01:09Z |
javascript
| 2017-11-05T14:46:46Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 11,454 |
["package.json", "scripts/rollup/build.js", "yarn.lock"]
|
Improve formatting of errors when building React
|
Currently when the build fails (e.g. due to a missing named export), the message is pretty ugly:
<img width="1033" alt="screen shot 2017-11-04 at 13 12 31" src="https://user-images.githubusercontent.com/810438/32405680-d618d2e6-c161-11e7-9b91-e0241c7c3eed.png">
It would be nice to display the relevant line inline with syntax highlighting, sort of like Babel does it. Maybe Babel's `babel-code-frame` could be helpful (not sure).
To reproduce, fail the build by removing some `export` somewhere and run `yarn build`. (Tip: `yarn build core --type=NODE_DEV` is the fastest way to build just `packages/react/*` and that might be enough to reproduce and iterate on the error message.)
The relevant logic for displaying build errors is probably somewhere in `scripts/rollup/build.js`.
**If you'd like to work on this please comment here, so we don't have many people working on the same thing.**
|
https://github.com/facebook/react/issues/11454
|
https://github.com/facebook/react/pull/11456
|
aa0b7418c17847552971441337c51d71f2ea5fc9
|
962042f82713b7d8fc61165cf4480801cbbd1f67
| 2017-11-04T13:14:57Z |
javascript
| 2017-11-19T14:23:33Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 11,444 |
["package.json", "scripts/tasks/linc.js"]
|
Fix `yarn linc` command on Windows
|
It currently fails on Windows because it depends on grep:
https://github.com/facebook/react/blob/61d35ce1f3f9c0706ed3630a2650d578497ae8c2/package.json#L101
We should find a way to make it do the same in a cross platform way (e.g. from a Node script).
|
https://github.com/facebook/react/issues/11444
|
https://github.com/facebook/react/pull/11447
|
366600d0b2b99ece8cd03d60e2a5454a02857502
|
743201387246d0cde523700c151550786f0afc2e
| 2017-11-03T18:33:45Z |
javascript
| 2017-11-04T18:09:28Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 11,443 |
["scripts/rollup/validate/index.js"]
|
Ensure linted bundles exist
|
I added validation of built bundles in https://github.com/facebook/react/pull/11432.
However I actually messed it up, and ESLint wasn't running on any files because of invalid arguments. I fixed that here: https://github.com/facebook/react/commit/61d35ce1f3f9c0706ed3630a2650d578497ae8c2#diff-edcddc0e0143d5a5a93d0951fd599794
We should fail the build if the glob patterns it's searching for don't exist.
|
https://github.com/facebook/react/issues/11443
|
https://github.com/facebook/react/pull/11452
|
028691b43d74d009c394cde16072d2b468dd88ef
|
da86a4553b2f8c71665b7d80e3405c1971451c2b
| 2017-11-03T18:17:10Z |
javascript
| 2017-11-04T12:57:48Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 11,441 |
["packages/react-test-renderer/src/ReactShallowRenderer.js", "packages/react-test-renderer/src/__tests__/ReactShallowRenderer-test.js"]
|
Remove support for cloneElement(getMountedInstance()) in shallow renderer
|
There was a test added as part of shallow renderer rewrite (#9426) in https://github.com/facebook/react/pull/9426/commits/fc1c6f913cb74a2cb1ce1cfe6ba91b431786c4ac that I don’t understand.
It tests that you can take an *instance*, call `cloneElement` on it, and pass the result to the shallow renderer. If we are doing this internally this seems like a misunderstanding of how the API should work, and we need to fix those callsites to read `getRenderOutput()` instead of `getMountedInstance()`.
You can see the test is wrong because if you change it to render to a *different* shallow renderer like this
```js
const shallowRenderer = createRenderer();
let result = shallowRenderer.render(<SimpleComponent foo="foo" />);
expect(result).toEqual(<div>foo:bar</div>);
const instance = shallowRenderer.getMountedInstance();
const shallowRenderer2 = createRenderer();
const cloned = React.cloneElement(instance, {foo: 'baz'});
result = shallowRenderer2.render(cloned);
expect(result).toEqual(<div>baz:bar</div>);
```
it will fail with
```
TypeError: Cannot read property 'prototype' of undefined
```
The argument we’re passing in this case is **not** a proper React element and doesn’t actually have `element.type`. It just mostly “happens” to work because both React elements and ShallowRenderer instances have `.props` on them, and code almost doesn’t read anything else in the update path. This [is an example hack](https://github.com/facebook/react/blob/779d23f72e7f2c763f7a3e3373a4134c54356fd6/packages/react-test-renderer/src/ReactShallowRenderer.js#L147-L148) that we needed because of this.
I think we should just delete this test, and fix the internal callsites. It was never intentionally supported, and prevents me from doing some other fixes in the shalllow renderer.
|
https://github.com/facebook/react/issues/11441
|
https://github.com/facebook/react/pull/11442
|
e8a382343aac011173a8eb68607e904d7ea26e40
|
fd47129ce6e8af37004686262bd27fa787af37fc
| 2017-11-03T14:22:34Z |
javascript
| 2017-11-03T16:16:56Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 11,435 |
["scripts/release/build-commands/add-git-tag.js", "scripts/release/build-commands/print-post-build-summary.js", "scripts/release/build-commands/update-package-versions.js", "scripts/release/build.js", "scripts/release/publish-commands/publish-to-npm.js"]
|
Release script follow-up items
|
Today I used the [new release script](https://github.com/facebook/react/tree/master/scripts/release) for the first time to publish `16.1.0-beta`. Here are a couple of things I noticed that I want to follow-up on:
### PR #11434
- [x] Auto-install (or update) release script Yarn deps before running
### PR #11437
- [x] `publish` script should `git tag -a ${version}`. I manually tagged the `16.1.0-beta` release.
- [X] The post-build test instructions should show fully-qualified paths to prevent you from accidentally verifying the wrong build (eg the checkout the release script is running from rather than the `-p` path)
- [x] `publish-to-npm` check should handle pre-1.0 packages. (This one failed for the `0.2.0` reconciler release b'c it expected it to be 16.1.0-beta.) For that matter, it should have probably been `0.2.0-beta`
- [x] When publishing a stable release (`@latest`) also update the next release to point to it.
|
https://github.com/facebook/react/issues/11435
|
https://github.com/facebook/react/pull/11437
|
43a1e0d084ce4491e7c61bb2b6ae9c768a1f4fad
|
af08b5cbcaf4d3e3ad965a9165e41688733a7771
| 2017-11-02T22:57:15Z |
javascript
| 2017-11-03T20:19:32Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 11,423 |
["packages/react-dom/src/__tests__/ReactDOMServerIntegrationElements-test.js", "packages/react-dom/src/__tests__/ReactServerRenderingHydration.js", "packages/react-dom/src/client/ReactDOMHostConfig.js"]
|
Ignore <noscript> content on the client and don't warn about mismatches
|
React 16.0.0 with SSR & [lazysizes](https://github.com/aFarkas/lazysizes) 4.0.1.
Trying to use the ["the noscript pattern"](https://github.com/aFarkas/lazysizes#the-noscript-pattern) to lazy load images with [lazysizes](https://github.com/aFarkas/lazysizes) but I'm seeing this:
> Warning: Expected server HTML to contain a matching <img> in <noscript>.
Image component render method:
```js
render () {
const { cdn, url, width, height } = this.props
if (!url) return null
const noScriptImgProps = {
src: `${cdn}${url}`,
className: classNames('product-image'),
width,
height
}
const imgProps = {
'data-src': `${cdn}${url}`,
className: classNames('product-image', 'lazyload'),
width,
height
}
return (
<span>
<noscript>
<img {...noScriptImgProps} />
</noscript>
<img {...imgProps} />
</span>
)
}
```
Does React have an issue with `noscript` tags..?
|
https://github.com/facebook/react/issues/11423
|
https://github.com/facebook/react/pull/13537
|
8a1e3962ab189b99b1593d8431feabcb4a21211b
|
0b74e95d7b7fabbd0cf502d0f3254a607ed30478
| 2017-11-01T13:51:13Z |
javascript
| 2018-09-03T16:17:53Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 11,419 |
["packages/react-dom/src/shared/HTMLDOMPropertyConfig.js"]
|
Capture attribute only supports boolean values
|
This is a follow-up to #4784.
The `capture` attribute on `input` elements is currently supported, but it is configured as a `HAS_BOOLEAN_VALUE` attribute: https://github.com/facebook/react/blob/1eed302d346bfb8f5db8d88b0e7096d8999d3548/packages/react-dom/src/shared/HTMLDOMPropertyConfig.js#L33
This causes
```
<input type="file" capture="user" />
```
to render as
```
<input type="file" capture />
```
The `capture` attribute supports strings though: https://www.w3.org/TR/html-media-capture/#the-capture-attribute
This is very important for being able to configure the default camera selected on mobile devices.
|
https://github.com/facebook/react/issues/11419
|
https://github.com/facebook/react/pull/11424
|
acabf112454e5545205da013266d8529599a2a82
|
2fe3494f0d6d42f8a0a4dd7a196d01a191b93f90
| 2017-10-31T22:50:05Z |
javascript
| 2017-11-12T14:29:27Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 11,379 |
["packages/react-dom/src/events/SelectEventPlugin.js", "packages/react-dom/src/events/__tests__/SelectEventPlugin-test.js"]
|
onDrop breaks onSelect event
|
**Do you want to request a *feature* or report a *bug*?**
Bug
**What is the current behavior?**
After an `onDrop` event has fired inside an "editable" element (input, textarea, contentEditable), the `onSelect` event on that element will no longer fire when changing selection using the keyboard.
As you can see in this gif below, the onSelect event gets called each time I move the cursor around (using the arrow keys). But once the `onDrop` event has fired, `onSelect` will no longer fire when I move the cursor around (also using the keys):
[Here's the jsfiddle](https://jsfiddle.net/owmL8n6u/1/) I used to record the video above.
**What is the expected behavior?**
The `onSelect` event should always fire when selection changes.
**Which versions of React, and which browser / OS are affected by this issue? Did this work in previous versions of React?**
React 16 / Chrome 62
|
https://github.com/facebook/react/issues/11379
|
https://github.com/facebook/react/pull/13422
|
146c9fb30763f84d85782dda82902f8513e70867
|
d14e443d6e97f6f4ca80d2e9a80300690b71e3c4
| 2017-10-26T22:46:34Z |
javascript
| 2018-08-16T23:22:16Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 11,348 |
["scripts/error-codes/replace-invariant-error-codes.js"]
|
Invariants that don't have error codes yet are missing in production builds
|
See my comment in https://github.com/facebook/react/pull/11291#issuecomment-338974574.
I verified https://github.com/facebook/react/pull/11291 introduced this regression.
|
https://github.com/facebook/react/issues/11348
|
https://github.com/facebook/react/pull/11350
|
d5f26b041e6dfce97e80176a4e278225fc8e47c7
|
793a28d69895c88d08ac1165e403125502df1f62
| 2017-10-24T12:35:22Z |
javascript
| 2017-10-24T16:11:03Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 11,310 |
["packages/react-reconciler/README.md"]
|
Example links in react-reconciler README are broken
|
Reporting this since it says _"If these links break please file an issue and we’ll fix them. They intentionally link to the latest versions since the API is still evolving."_ and knowing the place of the most up-to-date examples would help me a lot.
The issue template does not really fit here but I'm trying to follow it anyway.
-------------
**Do you want to request a *feature* or report a *bug*?**
Bug (In a README)
**What is the current behavior?**
In https://github.com/facebook/react/tree/master/packages/react-reconciler there are 3 links to reconciler configuration examples of `react-dom`, `react-art` and `react-native`.
All 3 links are broken
https://github.com/facebook/react/blob/master/packages/react-native-renderer/src/ReactNativeFiberRenderer.js
**If the current behavior is a bug, please provide the steps to reproduce and if possible a minimal demo of the problem via https://jsfiddle.net or similar (template for React 16: https://jsfiddle.net/Luktwrdm/, template for React 15: https://jsfiddle.net/hmbg7e9w/).**
Just visit the [`react-reconciler/README.md`](https://github.com/facebook/react/blob/master/packages/react-reconciler/README.md)
**What is the expected behavior?**
Correct links
I wanted to fix these in a PR but could only find an appropriate replacement for react-native here: https://github.com/facebook/react/blob/master/packages/react-native-renderer/src/ReactNativeFiberRenderer.js
**Which versions of React, and which browser / OS are affected by this issue? Did this work in previous versions of React?**
Issue exists on `master`-branch
|
https://github.com/facebook/react/issues/11310
|
https://github.com/facebook/react/pull/11312
|
97776fb6d211f3bcc363787d29c2824234e524f7
|
69fcc81ce4896ac0f0790a3bb4a7a842c349e447
| 2017-10-21T12:54:51Z |
javascript
| 2017-10-21T14:13:54Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 11,278 |
["scripts/release/build-commands/update-package-versions.js"]
|
React Test Renderer's Peer dependency keeps getting bumped
|
I noticed that react-test-renderer's peer dependency for React gets bumped with every version update (or at least that's what happens in version 15. v16 is still 16.0.0). Is this required? Could the peer dependency for version 16 not stay at ^16.0.0 even after the versions of each package get bumped to say 16.1?
I ask because this makes it hard for libraries to pull in react-test-renderer so that their end users (who do not use it directly) don't have to pull in a package they never use directly.
The specific example I have in mind is enzyme. For version 15, it has a peer dependency on `react@^15.5.0`, and a dependency on `react-test-renderer@^15.5.0`. if I have react 15.5 installed and enzyme tries to install the latest react-test-renderer, then it will pull in 15.6.2, which requires `react@^15.6.2`, which 15.5 does not satisfy.
Until recently, they made the test renderer a peer dependency, which was very annoying as a user of the library because you should not have to manually install a package that you don't use directly. We changed that to be a regular dependency, but are now noticing that the peer dependency the test-renderer has on react gets bumped every version change, which causes the above problem.
Is it not possible to leave the peer dependency within the major version? At least for v16 going forward? Or is there another way a library can pull in the test renderer without forcing the user to do it?
|
https://github.com/facebook/react/issues/11278
|
https://github.com/facebook/react/pull/11292
|
845b1afdc5ee82dd33751270fbde3cb8def4070a
|
fa5adb3bc44fae7ee97ad089f3e290be7153ed07
| 2017-10-19T04:08:49Z |
javascript
| 2017-10-20T15:53:15Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 11,254 |
["fixtures/art/VectorWidget.js", "packages/react-art/npm/Circle.js", "packages/react-art/npm/Rectangle.js", "packages/react-art/npm/Wedge.js", "packages/react-art/package.json"]
|
react-art on npm is missing the /lib/ folder
|
It seems like those were intended to be public APIs (https://github.com/facebook/react/pull/10629) but they’re not published on npm because `lib` folder is not in the `"files"` whitelist.
|
https://github.com/facebook/react/issues/11254
|
https://github.com/facebook/react/pull/11343
|
0ebcbd4f1ed67f6d7ca2ead61f1e63e0700d2c6b
|
f6c60dcbcd11c756e16b1f0e50c18ca4dc2e3953
| 2017-10-17T18:22:14Z |
javascript
| 2017-10-23T18:18:11Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 11,236 |
["packages/react-test-renderer/src/ReactShallowRenderer.js", "packages/react-test-renderer/src/__tests__/ReactShallowRenderer-test.js"]
|
Shallow test renderer calls shouldComponentUpdate on forceUpdate
|
There is no extra guard or clause in `enqueueForceUpdate` when calling `render` which means the `render` always calls `shouldComponentUpdate` which can block even forced re-render during tests.
This for example [breaks shallow rendering in Enzyme when using MobX and observables](https://github.com/airbnb/enzyme/issues/1267).
Minimal test case:
```javascript
import React from 'react'
import ShallowRenderer from 'react-test-renderer/shallow';
const renderer = new ShallowRenderer();
class TestComponent extends React.Component {
render() { return <span/> }
}
describe('forceUpdate', () => {
it('does not call shouldComponentUpdate', () => {
let called = 0
TestComponent.prototype.shouldComponentUpdate = () => called += 1
const component = renderer.render(<TestComponent/>);
renderer._instance.forceUpdate()
expect(called).toBe(0) // <- this throws because called === 1
})
})
```
|
https://github.com/facebook/react/issues/11236
|
https://github.com/facebook/react/pull/11439
|
c2b68dae640f87335fe224b7903e3448b50def5b
|
779d23f72e7f2c763f7a3e3373a4134c54356fd6
| 2017-10-16T15:10:37Z |
javascript
| 2017-11-03T11:03:06Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 11,175 |
["packages/react-dom/src/server/ReactPartialRenderer.js"]
|
Flow-ify ReactPartialRenderer.js
|
It doesn’t have a `@flow` annotation. We should start Flow-ifying it to avoid bugs like https://github.com/facebook/react/pull/11173.
@iamdustan Wanna take it?
|
https://github.com/facebook/react/issues/11175
|
https://github.com/facebook/react/pull/11251
|
b0bb1b46aac691766b39d8740472293126f84289
|
80849fd1c6c7a75d6aa862c53eb2119df706e82d
| 2017-10-10T15:33:56Z |
javascript
| 2017-10-26T16:59:29Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 11,172 |
["fixtures/dom/src/components/fixtures/textareas/index.js", "src/renderers/dom/fiber/ReactDOMFiberComponent.js"]
|
React 16 does not render text area placeholder initially in IE11 Win10
|
**Bug**
**What is the current behavior?**
IE11 Windows 10 text area placeholder not rendering initially.
[https://jsfiddle.net/fztx6vqp/](https://jsfiddle.net/fztx6vqp/)
**What is the expected behavior?**
Placeholder should be visible on initial render.
**Which versions of React, and which browser / OS are affected by this issue? Did this work in previous versions of React?**
16.0.0, IE11, Windows 10
Appears to be a regression in 16.0.0 - #6731
|
https://github.com/facebook/react/issues/11172
|
https://github.com/facebook/react/pull/11177
|
10320a5331d64fe83283d05c64c517402fce3512
|
309fa6c871362cfa8b2b94a7149585b5f599b73d
| 2017-10-10T13:34:55Z |
javascript
| 2017-10-10T16:49:59Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 11,152 |
["src/renderers/dom/shared/eventPlugins/__tests__/EnterLeaveEventPlugin-test.js", "src/renderers/shared/shared/ReactTreeTraversal.js"]
|
onMouseEnter triggered on children in React 16
|
Hi, I'm having an issue with `onMouseEnter` being triggered on children of a component. *It only occurs with React 16*.
I've created a [CodePen that shows the issue](https://codepen.io/testprofile75/pen/VMXXdb).
Simplified structure is as below:
```
<Card/> :
–––––––––––––––––––––––––––––
| |
| <Cover/> |
| |
------------------------------
| |
| <div class="text" /> |
| |
–––––––––––––––––––––––––––––
```
`<Card />` is a component that needs to capture `onMouseMove` events (and trigger a `render`) and `<Cover/>` is a simple component that waits for an image (passed as a `prop`) to be loaded before displaying it (using `onload` on an `Image` created in the `constructor`).
If you open the console, you'll see that `onMouseEnter` is triggered when hovering `<Card />` (normal behavior), but also when hovering the `.text` div (which is not supposed to happen).
If you remove the whole `onload` logic from the `<Cover/>` element, then `onMouseEnter` works as intended.
It's possible that I'm coding this in a way that isn't appropriate, but it does look like a bug anyway :)
|
https://github.com/facebook/react/issues/11152
|
https://github.com/facebook/react/pull/11164
|
08cbc257bdd75862c88ab1cef6ca2ac410cee8df
|
80596c2c92dce264f8572038730ce7aef647c5fa
| 2017-10-09T09:57:11Z |
javascript
| 2017-10-09T19:30:03Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 11,132 |
["fixtures/dom/src/components/fixtures/input-change-events/RangeKeyboardFixture.js"]
|
DOM Test Fixtures: Unable to Load React 16
|
As far as I can tell, the path to the UMD version of React 16 changed, which makes it impossible to test the latest version of React through the test fixtures. We probably need to add a condition to the react loader to change the path above a certain version:
https://github.com/facebook/react/blob/master/fixtures/dom/src/react-loader.js#L45-L46
|
https://github.com/facebook/react/issues/11132
|
https://github.com/facebook/react/pull/11133
|
7ec46e0fb0c8a865f0a3c87d563ceb46b05310ef
|
ea507f163b0b1b32446ed4dd4454580ed6d68f5c
| 2017-10-06T11:15:15Z |
javascript
| 2017-10-06T14:45:57Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 11,111 |
["scripts/rollup/bundles.js"]
|
Consider providing a UMD build for TestUtils again
|
I noticed we stopped providing it, but a bunch of fiddles for old bugs use it.
I wonder if we should add it back, e.g. as `react-dom-test-utils.development.js`.
Let’s hear what people have to say.
|
https://github.com/facebook/react/issues/11111
|
https://github.com/facebook/react/pull/11599
|
669a70dab789d58a7c26ec8b4bb897eef37466fa
|
adcf980333bf7365b0faa6e5510fc7b7362b70bf
| 2017-10-05T13:53:55Z |
javascript
| 2017-11-20T17:41:01Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 11,103 |
["src/renderers/dom/fiber/ReactDOMFiberComponent.js", "src/renderers/dom/shared/__tests__/ReactDOMServerIntegration-test.js"]
|
SSR: ReactDOM client and server handling newlines differently causing mismatch warnings
|
Bug report:
I noticed a "Warning: Text content did not match" warning on certain pages of my isomorphic React app, using React and ReactDOM 16.0.0.
After boiling it down to a simple test case it seems that different kinds of newlines are handled differently, some failing to match. If I do:
let test = 'foo\nbar';
...
<div>{test}</div>
Everything is fine, no mismatch error and the words are separated in the browser.
However if I do:
let test = 'foo\rbar';
I get:
```
Warning: Text content did not match. Server: "foo
bar" Client: "foo
bar"
```
Except in the JS error console the second foobar reads as one word. This seems to be a Chrome issue, mishandling single \r carriage return. Try `x = 'a\rb';` and it echoes that string to the console as one word. In the browser I see "foobar" as one word.
More importantly though if I do Windows newlines:
let x = 'foo\r\nbar';
The result in my app is:
```
Warning: Text content did not match. Server: "foo
bar" Client: "foo
bar"
```
Which definitely seems like a bug in React's reconciliation logic.
Setup:
Full list of NPM deps:
```
"dependencies": {
"babel-plugin-transform-object-rest-spread": "6.26.0",
"babel-preset-es2015": "6.24.1",
"babel-preset-react": "6.24.1",
"babel-register": "6.26.0",
"express": "4.15.3",
"json-bigint": "0.2.3",
"moment": "2.18.1",
"pm2": "2.7.1",
"react": "16.0.0",
"react-dom": "16.0.0",
"request": "2.81.0",
"request-promise-native": "1.0.4"
},
```
This is a Node/Express app on Node 8.4.0. Browser is macOS Chrome 61.0.3163.100
Workaround:
Preprocess any text that's user-uploaded content to normalize the newlines.
|
https://github.com/facebook/react/issues/11103
|
https://github.com/facebook/react/pull/11119
|
32ec797274ae2b956bab45aa335be50026ce1209
|
44c32fc26891867f2a35fb3c9bbedbedd148ccbb
| 2017-10-05T00:58:59Z |
javascript
| 2017-10-05T19:59:43Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 11,036 |
["www/gatsby-node.js", "www/src/layouts/index.js"]
|
Warning Documentation 404 Not Found
|
**Do you want to request a *feature* or report a *bug*?**
Bug
**What is the current behavior?**
The documentation for [warnings](https://github.com/facebook/react/tree/master/docs/warnings) returns a 404
**If the current behavior is a bug, please provide the steps to reproduce and if possible a minimal demo of the problem via https://jsfiddle.net or similar (template: https://jsfiddle.net/ebsrpraL/).**
https://cl.ly/080p0f1t400y
**What is the expected behavior?**
Redirect to correct documentation.
~~**Which versions of React, and which browser / OS are affected by this issue? Did this work in previous versions of React?**~~
---
The workaround is to look at the cached version from a google search or check for markdown [here](https://github.com/facebook/react/tree/master/docs/warnings).
|
https://github.com/facebook/react/issues/11036
|
https://github.com/facebook/react/pull/11038
|
4c8d248757cb18b9251411e0c3a2d29e951386d5
|
bbb272f3aae322bf2b13db862356d78957bafac7
| 2017-10-02T18:50:18Z |
javascript
| 2017-10-02T19:12:32Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 11,023 |
["src/renderers/dom/fiber/ReactDOMFiberComponent.js"]
|
[Feature request] Disable the "unrecognized tag" warning in React 16 but only for specific tags.
|
**Do you want to request a *feature* or report a *bug*?**
*Feature*
**What is the current behavior?**
In Firefox the `dialog` element is not implemented. We use a polyfill to be able to support it. Everything works well. But React warns that the `dialog` element is unrecognized, and this on each page load, which is quite useless (for this specific case).
I would like the possibility to disable the warning but only for some elements. So the utility of this warning remains when the developer mistype the component name.
**What is the expected behavior?**
Once we specified which tag are "accepted/normal" React will not warn on specified unrecognized tags.
**Which versions of React, and which browser / OS are affected by this issue? Did this work in previous versions of React?**
React 16
Firefox 57
In previous versions there was no warning.
Is this clear enough? Do you need any other information?
Cheers
|
https://github.com/facebook/react/issues/11023
|
https://github.com/facebook/react/pull/11035
|
9d4c2dfaac682c546bc746ed257b408b4e862f3d
|
223396ad0a63680a77b928712b34914ebc1e2788
| 2017-10-02T10:03:13Z |
javascript
| 2017-10-02T17:59:37Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 11,019 |
["packages/react-dom/src/__tests__/ReactTestUtils-test.js", "packages/react-dom/src/test-utils/ReactTestUtils.js"]
|
Deprecate and remove TestUtils.mockComponent
|
It is a Jest-ism that shouldn't be in our open source package.
People using Jest can implement a similar pattern directly.
|
https://github.com/facebook/react/issues/11019
|
https://github.com/facebook/react/pull/13193
|
afd46490d0b384ddf3f67f9a8fc857e32cdecfe3
|
d64d1ddb570a4fb7d748f4c500ac083886c2f0ec
| 2017-10-02T00:06:18Z |
javascript
| 2018-07-11T17:18:49Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 11,018 |
["scripts/prettier/index.js", "www/static/js/jsfiddle-integration-babel.js", "www/static/js/jsfiddle-integration.js"]
|
Old fiddles are broken after the new domain/website
|
http://fb.me/react-js-fiddle-integration.js no longer works, and this causes all old fiddles like http://jsfiddle.net/haa2euy9/1/ to break. We should fix it.
|
https://github.com/facebook/react/issues/11018
|
https://github.com/facebook/react/pull/11030
|
641ade94bd2887ad30186e856ade750a5f5f2d21
|
2fdb84d999d4041b2a0fd890121d949c1a672772
| 2017-10-01T23:49:11Z |
javascript
| 2017-10-02T15:48:59Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 11,017 |
["fixtures/dom/src/components/Header.js", "fixtures/dom/src/components/fixtures/custom-elements/index.js", "fixtures/dom/src/components/fixtures/index.js", "src/__mocks__/vendor/third_party/WebComponents.js", "src/renderers/dom/fiber/ReactDOMFiberEntry.js", "src/renderers/dom/shared/__tests__/ReactMount-test.js"]
|
Fiber cannot render to DocumentFragment/ShadowRoot
|
**Do you want to request a *feature* or report a *bug*?**
bug
**What is the current behavior?**
React 15.x was able to render within a DocumentFragment/shadowRoot, not possible with Fiber anymore
**DEMO**
https://www.webpackbin.com/bins/-KvPFG7-HGfQ34IgUxQt
```js
import React, {Component} from 'react'
import {render} from 'react-dom'
class MyComponent extends Component{
render(){
return <div>Hello From React !</div>
}
}
class MyElement extends HTMLElement {
static is = 'my-element'
constructor(){
super()
const shadowRoot = this.attachShadow({mode:'open'})
try {
// this fails
render(<MyComponent/>, shadowRoot)
} catch (e){
console.error(e)
shadowRoot.innerHTML = `
<div style="color:red;">${e}</div>
<div>see console log for stack trace</div>
`
}
}
connectedCallback(){
console.log('MyElement mounted!')
}
}
customElements.define(MyElement.is, MyElement)
```
related -> https://github.com/skatejs/renderer-react/issues/3
**What is the expected behavior?**
be able to render to shadow root like previous versions of React
**Which versions of React, and which browser / OS are affected by this issue? Did this work in previous versions of React?**
React 16
Chrome
|
https://github.com/facebook/react/issues/11017
|
https://github.com/facebook/react/pull/11037
|
bbb272f3aae322bf2b13db862356d78957bafac7
|
8b4ec79d4f4fea7fcea06ed58ec5329fbc64bafe
| 2017-10-01T23:33:20Z |
javascript
| 2017-10-02T19:50:14Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 11,010 |
["fixtures/dom/src/components/fixtures/selects/index.js", "src/renderers/dom/fiber/wrappers/ReactDOMFiberSelect.js"]
|
Reset of select ignores defaultValue
|
**Do you want to request a *feature* or report a *bug*?**
I report a bug.
**What is the current behavior?**
When resetting a form which contains a select with a defaultValue, the selected option becomes the first option element in the list, not the one with value=defaultValue.
**If the current behavior is a bug, please provide the steps to reproduce and if possible a minimal demo of the problem via https://jsfiddle.net or similar (template: https://jsfiddle.net/ebsrpraL/).**
I have put a piece of code here :
https://codepen.io/zehiro/pen/YrxZWw?editors=1111#0
Click on the reset button, which just does a form.reset() on the form.
**What is the expected behavior?**
I expect the select to select the option with value=defaultValue (like applying a form.reset() on a `<input/>`
**Which versions of React, and which browser / OS are affected by this issue? Did this work in previous versions of React?**
React-DOM 15.6.2.
It seems to be browser independent (tried with Chromium 61 and firefox 55 for Fedora).
I didn't try this on prior version of React.
|
https://github.com/facebook/react/issues/11010
|
https://github.com/facebook/react/pull/11057
|
b2101cb3281bff54af8527d6b3b7d78371395e37
|
5c6cb597f9a4e6eef4d1461b2af234498e52ac7f
| 2017-10-01T14:57:47Z |
javascript
| 2017-10-10T15:54:59Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 10,993 |
["fixtures/ssr/src/components/Chrome.js", "src/renderers/dom/fiber/ReactDOMFiberComponent.js"]
|
React 16 fails to rehydrate noscripts
|
**Do you want to request a *feature* or report a *bug*?**
Bug.
**What is the current behavior?**
As of React 16, `<noscript dangerouslySetInnerHTML={...} />`s are not properly rehydrated. Equivalent code works in React 15.
**If the current behavior is a bug, please provide the steps to reproduce and if possible a minimal demo of the problem via https://jsfiddle.net or similar (template: https://jsfiddle.net/ebsrpraL/).**
Consider the following server-side code:
```js
const React = require('react');
const ReactDOMServer = require('react-dom/server');
class Main extends React.Component {
render() {
const component = React.createElement('strong', {}, 'Hello!');
const contents = ReactDOMServer.renderToStaticMarkup(component);
console.log("--- Contents: " + contents);
return React.createElement('noscript', {
dangerouslySetInnerHTML: {__html: contents},
}, null);
}
}
function createApp() {
return React.createElement(Main, {}, null);
}
console.log(ReactDOMServer.renderToString(createApp()));
```
This code works fine in Reacts 15 and 16 alike; the outputs are similar:
```
// React 15
--- Contents: <strong>Hello!</strong>
<noscript data-reactroot="" data-reactid="1" data-react-checksum="1795300394"><strong>Hello!</strong></noscript>
// React 16
--- Contents: <strong>Hello!</strong>
<noscript data-reactroot=""><strong>Hello!</strong></noscript>
```
Now, consider the following application ([jsbin](https://jsbin.com/xonayepave/edit?html,output)):
```html
<!DOCTYPE html>
<html>
<body>
<!-- React 16 (exhibits bug) -->
<script crossorigin src="https://unpkg.com/react@16/umd/react.development.js"></script>
<script crossorigin src="https://unpkg.com/react-dom@16/umd/react-dom.development.js"></script>
<!-- React 15 (does not exhibit bug) -->
<!--
<script crossorigin src="https://cdnjs.cloudflare.com/ajax/libs/react/15.6.1/react.js"></script>
<script crossorigin src="https://cdnjs.cloudflare.com/ajax/libs/react/15.6.1/react-dom.js"></script>
-->
<div id="target"><noscript data-reactroot=""><strong>Hello!</strong></noscript></div>
<script>
class Main extends React.Component {
render() {
const contents = '<strong>Hello!</strong>';
console.log("--- Contents: " + contents);
return React.createElement('noscript', {
dangerouslySetInnerHTML: {__html: contents},
}, null);
}
}
function createApp() {
return React.createElement(Main, {}, null);
}
(ReactDOM.hydrate || ReactDOM.render)(
createApp(), document.getElementById('target'));
console.log("Initialized.");
</script>
</body>
</html>
```
In React 16, this displays the following in the console:
```
[log] --- Contents: <strong>Hello!</strong>
[err] Warning: Prop `dangerouslySetInnerHTML` did not match. Server: "<strong>Hello!</strong>" Client: "<strong>Hello!</strong>"
[log] Initialized.
```
If you swap the comment blocks in the HTML so that React 15 is used, the code works just fine: there are no warnings. (This is true regardless of whether you change the contents of `div#target` from the React 16 output to the React 15 output.)
The error message confuses me; it states that the server output is
```
<strong>Hello!</strong>
```
but this is not consistent with the output of the server-side code (neither the "Contents" nor the final markup emitted has HTML entities).
Please note that this is important because it breaks [the only known workaround][workaround] to #1252.
[workaround]: https://github.com/facebook/react/issues/1252#issuecomment-246161590
I believe that this may be specific to `noscript`s, because if one replaces all occurrences of `noscript` with `div` then the example works fine in React 16.
**What is the expected behavior?**
Rendering a component with `renderToString` and then rehydrating that component verbatim should not yield an error.
**Which versions of React, and which browser / OS are affected by this issue? Did this work in previous versions of React?**
Regression from 15 to 16. Tested in Chrome v61.
|
https://github.com/facebook/react/issues/10993
|
https://github.com/facebook/react/pull/11157
|
f42dfcdb943eb2c5299139820f626e8a81561165
|
e5db5302ac76e688ca044177d199435272c3ade1
| 2017-09-30T18:25:41Z |
javascript
| 2017-10-10T18:35:43Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 10,987 |
["fixtures/attribute-behavior/AttributeTableSnapshot.md", "src/renderers/dom/shared/HTMLDOMPropertyConfig.js", "src/renderers/dom/shared/__tests__/ReactDOMServerIntegration-test.js"]
|
<svg tabIndex="2"> doesn't work
|
See the original report in https://github.com/facebook/react/issues/10772#issuecomment-333242375.
The problem is that SVG is case sensitive, and so naïvely setting `tabIndex` and relying on insensitivity won't work with it.
The good news is it seems like `tabindex` is the only valid [SVG attribute](https://developer.mozilla.org/en-US/docs/Web/SVG/Attribute) with this problem. We just missed it because it applies both to HTML and SVG.
I think the fix is to add it back to the whitelist.
|
https://github.com/facebook/react/issues/10987
|
https://github.com/facebook/react/pull/11033
|
fbdd43fd70119c56ef06268a525168ebbb30e1f5
|
fbd6b9ded6f7c40bfde62c4881b364f54e73a026
| 2017-09-30T10:24:07Z |
javascript
| 2017-10-02T17:52:26Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 10,976 |
["www/src/components/CodeEditor/CodeEditor.js"]
|
React.org enable tab out of Live JSX Editor
|
The current React.org default behavior is that keyboard users will get to the code sections of the live editor and get stuck.
Can someone configure ignoreTabKey which makes the editor ignore tab key presses so that keyboard users can tab past the editor without getting stuck?
Chrome on Windows atm and cannot edit it to submit a fix. Sorry about that :(
|
https://github.com/facebook/react/issues/10976
|
https://github.com/facebook/react/pull/10992
|
25dbf03e66efa7bd3fe854df677e3c69bb45d5e2
|
53c1b8b0d7109748d95592ea4fdc50e490d9565c
| 2017-09-29T20:26:43Z |
javascript
| 2017-10-03T18:15:00Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 10,965 |
["www/src/css/algolia.css"]
|
[website] docsearch improvements
|
**What is the current behavior?**
the docsearch on the new page changed its design, but two features are different than in the previous version:
- the "cursor" no longer has a separate colour (visible when you use the arrow keys to select a result)
- there no longer is an Algolia logo / mention
**What is the expected behavior?**
- A slightly different background colour for the cursor
- A mention of "search by Algolia" or our logo, as which is required because it's a free plan.
Happy to open a PR for both of those issues, with whatever colour / solution we choose
cc @bvaughn and @gaearon
|
https://github.com/facebook/react/issues/10965
|
https://github.com/facebook/react/pull/11032
|
f6a79d1f1607fb5e50904b143edcf3801144d2ce
|
e367a4478913ca5bf52fa2260897718f7d16035b
| 2017-09-29T16:36:29Z |
javascript
| 2017-10-03T17:39:14Z |
closed
|
facebook/react
|
https://github.com/facebook/react
| 10,925 |
["www/src/prism-styles.js"]
|
[website] highlighted lines in code snippets don't have enough contrast
|
**Do you want to request a *feature* or report a *bug*?**
This is an accessibility issue with the new documentation website.
It's a clear-cut problem, but the solution is pretty open ended and ill defined. Might be difficult, and could use some design creativity.
**What is the current behavior?**
We are using [Prism.js](http://prismjs.com/) for the syntax highlighting styles of our code snippets. We also sometimes want to "highlight" a certain line, so we lighten the background of that line. With the lighter background, many of the font colors don't provide enough contrast.
**If the current behavior is a bug, please provide the steps to reproduce and if possible a minimal demo of the problem via https://jsfiddle.net or similar (template: https://jsfiddle.net/ebsrpraL/).**
You can verify this by [running the aXe a11y audit](https://www.deque.com/products/axe/) on this page: https://reactjs.org/docs/introducing-jsx.html
**What is the expected behavior?**
We need a way to highlight a line without losing so much contrast.
BUT choosing a different hue without making it lighter looks bad in my opinion, and I would rather not fiddle with the various font colors used in the syntax highlighting.
We should either collaborate with the Prism project to add a 'highlight' option with enough contrast, or find another library/system to use for syntax highlighting.
|
https://github.com/facebook/react/issues/10925
|
https://github.com/facebook/react/pull/10931
|
b43574067e3bf9bfbf9fd772bd82a010b8c88dda
|
7ae5ed0e7b087a202a8cfa24d7b4b5cde7c913e1
| 2017-09-28T17:51:32Z |
javascript
| 2017-09-30T14:54:20Z |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.