I'm trying to render the following component:
function deviceready() {
ReactDOM.render(
<div class="container">
<h1>Hello <small>World</small></h1>
<button class="button-primary" onclick="initFacebook()">Login with Facebook</button>
</div>,
document.getElementById('frame')
);
}
...With the reactify gulp plugin. However the attributes don't get compiled. class="container" and class="button-primary" and onclick="initFacebook()" do not appear in my generated HTML.
I'm aware that custom attributes are not supported but these are attributes defined by HTML spec. Any ideas what I'm doing wrong?
Notice the differences pointed out in Facebook's JSX in-depth article.
class should be className
Also notice the Event System; the prop names don't match up exactly with the HTML attribute names
onclick should be onClick
Related
I have a GatsbyJS site that I am working on where the main content source is a Wordpress install. One of the things I like to add to my sites is the ability to have placeholder areas in the site where I can control the content via the CMS. Usually I have a header_scripts area that goes at the bottom of the <head> tag, a body_scripts area that goes at the start of the <body> tag, and a footer_scripts area that goes at the bottom of the page <body>. With these three, I can usually integrate third-party add-ins pretty easily without having to do code deployments.
Sometimes I need to embed stylesheets, sometimes I need to embed script tags, and sometimes I need to throw in <meta> tags. Really the content could be anything. This data comes back as a raw string from my Wordpress GraphQL endpoint.
So now my question is, how do I get this content injected into my Gatsby site in the following places:
<html>
<head>
...
{header_scripts}
</head>
<body>
{body_scripts}
...
{footer_scripts}
</body>
</html>
I've found so far that I can just include the body_scripts and footer_scripts in a fairly regular manner in my Gatsby page template. In gatsby-node.js, I pass in the property values using the pageContext. It's kind of a bummer that they need to be wrapped in a <div /> tag, but they seem to work just fine.
import React from 'react'
export default class PageTemplate extends React.Component {
render = () => {
return (
<React.Fragment>
{this.props.pageContext.bodyScripts && (
<div dangerouslySetInnerHTML={{__html:this.props.pageContext.bodyScripts}} />
)}
{/* my page content here */}
{this.props.pageContext.footerScripts && (
<div dangerouslySetInnerHTML={{__html:this.props.pageContext.footerScripts}} />
)}
</React.Fragment>
)
}
}
Now for the real question. I am stumped on how to get the dynamic content from the header_scripts into the Gatsby server-side-rendering <head> tag. The closest thing I have found to being able to inject content into the head is to leverage the gatsby-ssr.js onRenderBody function. However, this seems to require pre-determined React component instances in order to function. I can't just pass it in plain raw string content and see the output in the page source:
export const onRenderBody = async ({
pathname,
setHeadComponents,
setHtmlAttributes,
setBodyAttributes,
setPreBodyComponents,
setPostBodyComponents,
setBodyProps
}, pluginOptions) => {
setHeadComponents(['<script>alert("hello");</script>'])
}
This results in an escaped string getting inserted into the <head> tag:
<html>
<head>
...
<script>alert("hello");</script>
</head>
<body>
...
</body>
</html>
I'm at a loss as to how to proceed. I can't just wrap my string in a <div /> tag like in the body because div tags can't go inside the head tag. I can't think of any head-capable HTML tags that would accept this kind of content.
The only idea I've had is to actually parse the string content into full React components. This seems daunting given the number of possible tags & formatting that I would need to support.
Am I going about this the wrong way? How can I get my arbitrary content into my Gatsby site's head tag?
It's a broad question and it will need some trials and errors to ensure that it's fully working without caveats in all scenarios but, among the things you've tried, you can add a few more options to the list to check which ones fit better.
Regarding the body_scripts and footer_scripts both can be inserted using the:
<div dangerouslySetInnerHTML={{__html:this.props.pageContext.footerScripts}} />
In any desired page or template. For the header_scripts and the meta tags (SEO), you can use the <Helmet> component. Basically, using this component, everything that is wrapped inside, it's becomes transpiled inside the <head> tag once compiled.
export default class PageTemplate extends React.Component {
render = () => {
return (
<React.Fragment>
<Helmet>
{this.props.pageContext.headerScripts && (
<div dangerouslySetInnerHTML={{__html:this.props.pageContext.headScripts}} />
)}
</Helmet>
{this.props.pageContext.bodyScripts && (
<div dangerouslySetInnerHTML={{__html:this.props.pageContext.bodyScripts}} />
)}
{/* my page content here */}
{this.props.pageContext.footerScripts && (
<div dangerouslySetInnerHTML={{__html:this.props.pageContext.footerScripts}} />
)}
</React.Fragment>
)
}
}
However, if the data comes from a CMS, it won't be available in the SSR yet, so, one easy thing you can do is to customize the outputted HTML (html.js) that Gatsby generates in each compilation. From the docs:
Customizing html.js is a workaround solution for when the use of the
appropriate APIs is not available in gatsby-ssr.js. Consider using
onRenderBody or onPreRenderHTML instead of the method above. As a
further consideration, customizing html.js is not supported within a
Gatsby Theme. Use the API methods mentioned instead.
Run:
cp .cache/default-html.js src/html.js
Or manually, copy the .cache/default-html.js file and paste it /src folder. There you can customize the final HTML.
I need create a component to render like this:
<div class="row">
<div class="col">
<div class="some another class">
<h3>{{title}}</title>
</div>
</div>
<div class="col">
<p>{{contentA}}</p>
</div>
<div class="row">
<div class="some wrapper">
<p>{{contentB}}</p>
<div>
</div>
<div>
It will receive the following data:
title - a string
contentA - a large string (paragraph), may contain
links (a href)
contentB - a large string or a complex HTML code
Here are the following methods I found. However, I'm not sure about which one is the recommended way by React.
Method 1
<Component title="a" contentA="some large content" contentB="another large content"/>
Pros: Easy to access different props in different sections of the component
Cons: Too complex to handle if the content as large HTML content
Method 2
<Component>
<title>a</title
<content>some large content</content>
<content>another large content</content>
</Component>
Pros: Much simpler code
Cons: Need to filter props.children and find each element to put in correct place in the component
This is not really an accurate statement:
Cons: Too complex to handle if the content as HTML contents
If a value can be held in a variable as a string, it can be used as an argument passed into your component. If you don't want to be bothered trying to escape the quotes inside your prop values, I would encourage you to ditch the <Component attr="someValue"/> syntax and instead use <Component attr={'someValue'}/>. This allows you to pass variables into your props. This also allows you to use template literals to pass values, like so:
<Component
title="a"
contentA={`
<html>
<head></head>
<body>
<p>Here is my super-big</p>
<p>HTML-infused content</p>
<p>I can even inject ${variables} in here!</p>
</body>
</html>
`}
contentB="another large content"
/>
If you feel like that starts to clutter-up the declaration of your component, you can instead set those values in a variable and then pass them very cleanly into the component, as such:
render() {
const bigHtmlContent = (
<html>
<head></head>
<body>
<p>Here is my super-big</p>
<p>HTML-infused content</p>
<p>I can even inject {variables} in here!</p>
</body>
</html>
);
return (
<Component
title="a"
contentA={bigHtmlContent}
contentB="another large content"
/>
);
}
Of course, you can move the definition out of the render() function altogether if that suits you better:
getBigHtmlContent() {
return (
<html>
<head></head>
<body>
<p>Here is my super-big</p>
<p>HTML-infused content</p>
<p>I can even inject {variables} in here!</p>
</body>
</html>
);
}
render() {
return (
<Component
title="a"
contentA={() => this.getBigHtmlContent()}
contentB="another large content"
/>
);
}
When I'm creating components, I tend to pass everything as props (as shown in the examples above). I only use children if I'm creating a higher-order component which, by design, is supposed to act as a container for certain types of children.
For example, in Material UI there is a <List> component which requires one-or-more <ListItem>s. In that case, I don't think it would make much sense to pass the <ListItems>s as props. They are (and should be) children of the <List> component.
React is all about creating granular components, the more granular you component tree is, you are doing it the more React way.
You say you get 3 things text,contentA and ContentB
Title: it's already granular.
content A: You say its a "a large string (paragraph), may container links (a href)"
, if its a paragraph it is granular but if its a collection of paragraphs, I think you should take that to a different component. Ex: ContentA component
Content B: This is clearly said that its a complex HTML code, that is what is React for, you need to break this complex HTML code into granular components which will make it more maintainable and it is a React best practice too.
have patience while breaking this into components, it's definitely worth it.
hope that helps.
I am new to react and could not grasp the concept between bsClass and className.
I try to implement a modified button style, like:
<Button bsClass="btn-custom" >Custom button</Button>
where it does not work when I substitute bsClass with className.
But in other part, using the same custom.css source, I implement:
<img src={logo} alt="logo" className="app-logo" /> and it works.
JSX attribute className is equivalent of HTML class. So the below JSX
<span className="app-logo">Logo</span>
will be equivalent to the below in HTML
<span class="app-logo">Logo</span>
As per bsClass is considered in
<Button bsClass="btn-custom" >Custom button</Button>
it is prop that is being passed on to the Button component in reactJS and that is what it will be using to set the className inside the component something like
<button className={this.props.bsClass}>{this.props.children}</button>
So it an attribute that is defined as a property by the react-bootstrap docs.
React's className is exactly equivalent to regular classes in CSS.
HTML/CSS:
<div class='red-text'>
Foo
</div>
React/JSX:
<div className='red-text'>
Foo
</div>
The above snippets of code do the exact same thing. The only reason we're stuck with using className in React (instead of class) is because "class" was already taken as a reserved keyword in JavaScript.
As the others have said, bsClass is a pre-defined class within the react-bootstrap package. Just like how the CSS-version of Bootstrap comes with its own styling, so, too, does react-bootstrap.
A practical difference. I you set bsClass to something other than what REACT-Bootstrap has as a default, you have to do your own css themeing of the button.
By adding a className="xx" you still get the default theme, but you can now add css styles for color, padding, etc, using your own .css
.xx {
magin-bottom: 2px
}
I'm trying to figure out how to make a component render its children.
So I can compile:
<my-component>
<div id="child"></div>
</my-component>
into something like this:
<div id="parent">
<!-- some component stuff -->
<div id="child"></div>
</div>
Is there something like ngTransclude in Angular.Dart?
AngularDart uses Shadow DOM in place of ngTransclude. This is one of the ways we are pushing Angular into the browser with emerging web standards.
Adding a <content> tag instead your component's template will cause the child element to be placed there.
e.g. in your example <my-component>'s template might look like:
<div id="parent">
<!-- some component stuff -->
<content></content>
</div>
For even more awesomeness, you can use Shadow DOM's content selectors as well to control which child elements are displayed.
I think some sample code can explain my purpose.
Some html code with angular:
<div ng-init="buttons=['add','edit','delete']">
<div show-result-as-text>
<button ng-repeat="button in buttons">{{button}}</button>
</div>
</div>
You can see there is a custom directive "show-result-as-text" which I want to define. It should render the inner html code with angular directives, then show them as text.
The final html should be:
<div ng-init="buttons=['add','edit','delete']">
<div show-result-as-text>
<button>add</button>
<button>edit</button>
<button>delete</button>
</div>
</div>
And when the buttons value changes, the escaped html should also be changed.
I've tried to write one myself, but failed after 2 hours of work.
UPDATE
A live demo: http://plnkr.co/edit/fpqeTJefd6ZwVFEbB1cw
The closest thing I could think of is exemplified here: http://jsfiddle.net/bmleite/5tRzM/
Basically it consists in hiding the src element and append a new element that will contain the outerHTML of each src child.
Note: I don't like the solution but it works, so I decided to share it...