Let's say I have a very simple app with the following render method.
render(){
return (<div style={{height:window.innerHeight}}/>);
}
Resulting in the following DOM:
<div data-reactroot="">
<div style="height:938px">
</div>
</div>
Now when I try to rehydrate the DOM I get a warning saying that style prop does not match but otherwise everyhing still works allright, except for the fact that the height of the div does not get updated to what it should be.
What would be the best way to either fix the warning or ignore it (as everything else work) and make react update the DOM element.
Note that:
- I think this worked fine w. React 15.x. Using 16.0.0 now;
- In this toy example I could obviously just use height:"100%" but in the real life case I do not have this luxury and have to calculate the dimensions in a way that depends ultimately on viewport size.
Did a McGyver fix here, reading the console for those warnings and then if I get those:
setTimeout(function(){
ReactDOM.unmountComponentAtNode(RoboReact.rootParent);
ReactDOM.render(RoboReact.rootElement,RoboReact.rootParent);
},100);
(doesn't seem to work w/o the setTimeout for some reason)
Related
I'm currently experimenting with React, and I've now run into an issue that I can't seem to solve.
In my application, I use a React library to handle hotkeys, these hotkeys have a scope, so when I want a certain set of hotkeys to be active in a div, I have to wrap that div with a <HotKeys> tag.
I have to toggle some of these divs, so I'll have something along the lines of
isActive ?
<HotKeys ...>
<div ...>...</div>
</HotKeys>
: <div ...>...</div>
I now need to figure out a way to focus the div when it's created. Pretty much every article on the web suggest something like this:
const focusRef = useRef(null);
useEffect(() => {
focusRef.current.focus()
})
return (
isActive ?
<HotKeys ...>
<div ref={focusRef} tabIndex={-1} ...>...</div>
</HotKeys>
: <div ...>...</div>
)
I've tried some variations, including having the div top level (without the <HotKeys> wrapping them), all to no avail.
When I print the focusRef object in the useEffect method, I do get the expected output, focusRef is correctly set and current is populated, but calling the focus method doesn't work. At one point I tried calling the focus method from a button and manually triggering it after the component had fully loaded, it seemed to work (document.activeElement was being changed), then for some reason it stopped working again. All this leads me to believe that somehow the component hasn't fully loaded, despite calling the useEffect hook, which, if I understand correctly, triggers when the element has rendered for the first time/after every change to state.
I'd really appreciate some help with this, since I basically started learning React yesterday.
You must to use an useCallback , because useRef don't notifies you when ref was created
https://reactjs.org/docs/hooks-reference.html#useref
I think I figured it out.
You were right about using the tabIndex but you needed to pass it in as a string like this:
tabIndex={"-1"}
When you first load it a dotted line box surrounds the div that has the ref attached.
check out this code sandbox:
https://codesandbox.io/s/nifty-wildflower-eqjtk?file=/src/App.js
grabbed from this accepted answer where they pass in a string:
Need to put focus on div in react
Update - JSFiddle demo here: https://jsfiddle.net/vb2ptmuo/11/
I'm playing around with an interesting effect I came across: https://tympanus.net/codrops/2019/08/07/image-trail-effects/
To start, I have a div containing a bunch of img elements. It dumps them into an array and then creates a trail effect from those images which follows the mouse. You kick this all off via new ImageTrail(".content"). But what if I have more than one set of images and I want to re-trigger it again with those, instead? Example:
<div class="content">
<img src="1.jpg">
<img src="2.jpg">
<img src="3.jpg">
</div>
<div class="content-2">
<img src="4.jpg">
<img src="5.jpg">
<img src="6.jpg">
</div>
Doing a second new ImageTrail(".content-2") does not replace the first set of images with the second set, even though the code reads to me like it should. You still just see the first set of images in the trail.
I'm also slightly concerned with performance if I'm instantiating the ImageTrail class twice (if that's even a thing), but this is wholly secondary to my main issue.
I feel like there's a simple solution but I'm missing it. Scroll down to the bottom of the demo for the commented code "This doesn't work"
The main cause is requestAnimationFrame as part of constructor and render method.
Across instances, for this use case, one instance should have the rendering control using requestAnimationFrame.
I did a trick here, for this use case.
Every time a new instance is created, it will cancel earlier instance's render request (done by requestAnimationFrame) by means of cancelAnimationFrame
Even i cancelled the request done in render method too.
Check out this jsfiddle for modified code : https://jsfiddle.net/rahultarafdar/mfboqy9g/45/
The Code that you posted has one major problem: it uses global variables to store state of the ImageTrail object. This is a pretty bad programming practice as it leads to unexpected behaviour and defeats the purpose of having a class in the first place.
The reason why only the first set of pictures is shown is because both instances of ImageTrail register a callback for the next animation frame with requestAnimationFrame. The callback (render) of the first instance is always called first, because it registered the listener first. This first render function updates the global mousePos variables. The second render function thinks the mouse has not moved, because lastMousePos and mousePos are the same.
Now to address your problem: first you should move all the state that is used by ImageTrail into the class itsef. That includes mousePos, cacheMousePos and lastMousePos. If you have done that successfully, you should get both imagesets diyplaying at the same time, if you have two instances. To activate/deactivate the rendering of ImageTrails, you could add an active attribute, which gets checked at every render call or you could implement methods that cancel/setup the animation frame for a specific instance. (For details how to use cancelAnimationFrame you could look into the answer from rahul or into the MDN)
Thought everything was going great. Works perfectly in Chrome, FF, Edge, even IE 11!
The parent component holds all the sate. I pass the bets object to the child component which calculates a count to pass to the grandchild component to display.
The parent state 'bets' is an object with the keys as an ID and the value as an object.
The parent state is correctly changing when I interact with the app. Why will only Safari not update when the parent state changes? (on iOS and MacOS)
Parent
class ParentComponent extends React.Component {
constructor(props) {
super(props);
this.state = {
bets: {}
};
}
}
Child
getBadgeCount = (league) => {
const bets = this.props.bets;
let count = 0;
Object.keys(bets).map((bet) => bets[bet].event.league === league && count++);
return count;
};
// ...
<ChildItem count={this.getBadgeCount(league)} />
GrandChild
class GrandChildComponent extends React.Component {
render() {
const { count } = this.props;
return (
<div>
<div>
{count > 0 && <div>{count}</div>}
</div>
</div>
);
}
}
I console.log the count inside the grandchild render and in the componentDidUpdate and it shows the right number. Is there something apple/safari specific with react I am missing?
In the case where there are variables being changed based on which you can switch between the styles (opacity: 1 or opacity: 0.99) You can try adding a key to the element which is not being updated in Safari/iOS.
<div key={new Date()} className={'myComponent'}>{Count}</div>
I ran into the same problem, this seems to work for now.
So I did solve the very bizarre issue after a few days.
What was happening was I guess the engine in the Safari browser was not re-rendering the little badge icon correctly. I would print out the value inside the DOM, OUTSIDE of the styled badge I was using, and the number would update as expected... Leading me to believe the issue was related to styles.
THEN after removing what I thought was the CSS causing the issue, I noticed that it looked like safari was 'partially' updating the value in the badge. It appears to have half re-rendered, and the previous number collides with the new value... Not sure if this is some obscure problem with the render engine?
After adding and removing all the CSS one by one, the issue remained so I decided to trick the browser to 'force' render with a simple calculation inside the grandchild where it was being rendered:
const safariRenderHack = { opacity: count % 2 ? 1 : 0.99 };
<div style={safariRenderHack}>{count}</div>
Not a great solution but oh well I guess it's fixed. ha
I just run into similar problem.
Safari didn't re-render some texts implemented as:
<span>{someValue}</span>
On 'someValue' field update, reactJS worked fine (element was requested to render) but Safari re-renders only area of new value (shorter than previous). UI glitches :-/
If I done anything to CSS via Developer tools, element has been rendered again and looks fine :-/
After some tries, I luckily used a 'display: block;' style property and it starts to re-rendering absolutely fine. Also 'display: inline-block;' will fix that problem too, if it is needed to be used.
I had a similar problem where a center positioned span using flexbox inside a div wasn't updating correctly in Safari.
Restyling this div to use display:block solved this rendering issue so this might be worth looking into if anyone runs into the same problem in the future.
P.S. Looking at the styling code now, making the change was actually a much cleaner solution as it reduced the amount of lines from 4 to 1 but this might not be the case for all scenarios of course.
before:
display: flex;
flex-direction: row;
justify-content: center;
align-items: center;
after: (is a div so display:block is already used)
text-align: center;
When using map function , React looks at the keys :
To identify which items have changed, are added, or are removed.
Solution :
Now, As you use map function u must Add Key to the div inside map .
Use a key variable that is connected to the logic behind clicking , so that when clicking the key , the variable is updated and so React will rerender this grandchild div .
As the following code :
<div key={Count} className={'myComponent'}>{Count}</div>
PS : I discourage to use Date as a key , because I f u are clicking quickly it glitches , while if u are connecting it with a state with the logic behind clicking , it won't .
this may help you:
will-change: transform;
I know it may be late but what helped me with Safari not re-rendering reused components was:
will-change: opacity;
You may have to experiment which element in your HTML hierarchy to apply this to.
It makes no sense because no opacity was being changed in my case but it did fix the problem. Thank you, Apple...
We also ran into this problem. So, I did a short research about it and got the gist as follows:
It works by providing the unique key to the element(key={new Date()} works bcz unique even with seconds) which is not rerendering properly say a description as text inside it that changes by button or an event.
So what I got as a gist basically, providing the individual key make new dom element to react during reconcilation process rather than changing same value inside the element so basically that element showing object is made unique instead of changing value inside that dom element so I think safari specific it is as it is holding state of the element rendering dynamically. So after providing the key prop you are basically rendering the unique different element(obviously with different text/object inside it). This basically utilizes here the concept of keys while we render a list in react component.
Inserting dynamic SVG content into the DOM does not work as expected, having the SVG element onload attribute (containing JavaScript) regarding: "setInterval()".
As noted in the search tags of this question; this is plain (valilla) JavaScript (not jQuery); here's a breakdown of the issue:
I have some SVG code (plain text) that gets inserted into a <div> as innerHTML
the SVG element has an onload attribute with some JavaScript inside it
the JavaScript contains setInterval(...) - (which does not work at all)
I grab the SVG element from the temporary div and return it as the result of a function.
this result is appended into some element in the live DOM (body)
the strange issue:
any other code inside that onload attribute works fine,
only setInterval & setTimeout is completely ignored
More info:
During runtime (start-up), the SVG code is grabbed from an existing embed element .getSVGDocument() (after it has loaded) and prepared as plain HTML which can just be used as a template to create many others from the same source-code. I'm not using cloneNode(true) -because: the interval is for animation (continuous slow & smooth rotation) - which could have a heavy impact on client-side resources, hence, I thought it best to grab the code and keep it as template - then remove the original from the DOM.
With all the above in mind, everything works fine:
The (new) SVG shows up on screen, all nice and dandy-like
When I console.log the (inline) SVG code that is used, all looks perfect
I get no errors, and there is no error handler that mutes errors (window.onerror == null)
The JavaScript (text) inside the SVG node's onload attribute works for things like: console.log(this) - (which produces the SVG element in the log) - however, as mentioned: setInterval() & setTimeout() is just simply ignored - without any error and no warning.
The following code is a very short example, and (regrettably) it works; however, in my production app it doesn't.
The code:
var html = '<svg xmlns="http://www.w3.org/2000/svg" version="1.1" onload="setInterval(function(){ console.log(\'testing\'); },500);">';
var temp = document.createElement('div'); temp.innerHTML = html;
var node = temp.getElementsByTagName('svg')[0];
document.body.appendChild(node);
If you test the above code in a new js file, it works; however, for the life of me I can't find the reason why it breaks in my app; as explained, it's quite simple really.
The question:
Does anyone know if there is some "gotcha" I'm not aware of regarding this? Maybe name-spacing?
If the source-code is required, I can load it up on JSFiddle, or CodePen -if required, but, it's a lot of code, and many files, which may not be necessary for publication.
I'm sure it's just something small; like, how timers register according to scope, and maybe how it's affected in .bind() ?
I'm really stuck with this, and I kinda need it working for a good impression for a job-interview; so if you know anything that could help, I would appreciate your input very much.
Thank you.
embedded content, onload attributes & the DOM
The following may help in related scenarios:
when targeting an asynchronous source, make sure the contentDocument or getSVGDocument() contains the resources you need to access. The window.onload, or the DOMContentLoaded event is relative to the current DOM, so it may help constructing your own listener->trigger for a cross-browser solution, because the contents you need may not be ready in a synchronous fashion.
the onload attribute/event is not triggered when inserting dynamic content that is not asynchronously loaded, but may fire under certain circumstances, so, again, a custom:
listen->trigger will solve that.
question specific
The question is directly related to the 2nd point above, and the answer is quite simple really:
in the "onload" attribute of said SVG, set a simple value as property of this like:
<svg onload="this.ready = true; setTinterval(...)"
in the constructor function, after the element was dynamically created, simply check if the svg-node's onload event was fired like so:
if (!svgNode.ready){ svgNode.onload(); }
If there is an error in your code, but no error is shown, make sure window.onerror is either null -or if it's a function, make sure it does NOT return true - else it may suppress errors and you'll have a hard time tracking down what's wrong.
Please feel free to either improve this answer, or comment and I'll improve it accordingly; however, better answers will be appreciated.
6 years later...
With vanilla JavaScript Web Components you can do:
<load-svg></load-svg>
<script>
customElements.define("load-svg", class extends HTMLElement {
connectedCallback() {
this.innerHTML = `<svg></svg>`;
setInterval(() => {
console.log("testing");
}, 500);
}
});
</script>
I'm trying to do some dynamic resizing using javascript. For part of this I need to report the current size of a container. I wrote a function to do this (below) and it works great in chrome... however in firefox the width does not get reported. I have had a read on the offsetWidth method and in theory it should work in firefox.
function currentBoxW() {
return document.getElementById("wrapper").offsetWidth;
}
All this reports is "NaN".
There is more in that function in the actual page, but all the additional code has been commented out for testing... this is the actual line that does the reporting.
Have I missed something or is there a smarter way to report width using JS and the DOM? Is there another way to get the width of a div?
Thanks.
You can have two problems:
Make sure you have explicity set the width of the element (wrapper) through CSS
Make sure javascript code is executing after document is ready.
window.onload() or something similar if you are using any other javascript libaries
If you can use jquery instead
you can easily get it using
$(document).ready(function(){
function currentBoxW() {
return $('#wrapper').width();
or
return $('#wrapper').innerWidth(); //depending on the requirement.
}
});