Adding content script will make pointer-events: none (chrome extension) [duplicate] - google-chrome-extension

I wrote a Google Chrome extension, which popups a dialog with an autocomplete field and it's own style, but there are some sites where my CSS gets totally broken, which doesn't look very nice.
I know about isolating styles with iFrames, but in Google Chrome extension there is no way to isolate my HTML and CSS in this way. Another method is to wrap all my stuff into a separated div with it's own id and relative styles for that id, and I do so, but it seems that it doesn't work on some sites with "hard" tags style overloading or "!important" directives in the CSS code.
So, I want to know is there any way to really isolate my styles in z convenient way or it's my bad carma to overload every little CSS property to fix one or another style issue for each site?
By the way: I set up my manifest to load all the things at the "document_end", but I see it's not being applied to the stylesheets which is every time loaded whenever the DOM is ready.

At the time of asking the question, your only option was to either use iframes, or stylesheets with a very high specificity and explicitly set all properties that might affect styles. The last method is very cumbersome, because there will always be some property that is overlooked by you. Consequently, the only usable method for isolating stylesheets was to use iframes.
The solution to this problem -isolation of styles without iframes- is Shadow DOM (since Chrome 25). You can find a tutorial at HTML5 Rocks. For a real-world Chrome extension that uses Shadow DOM to isolate styles, see Display #Anchors (source code here).

As I've recently gone through the gauntlet of this issue, I want to share some information I think is valuable.
First, Rob W's answer is correct. Shadow DOM is the correct solution to this problem. However, in my case not only did I need CSS isolation, I also needed JavaScript events. For example, what happens if the user clicks a button that lives within the isolated HTML? This gets really ugly with just Shadow DOM, but we have another Web Components technology, Custom Elements, to the rescue. Except that as of this writing there is a bug in chrome that prevents custom element in chrome extensions. See my questions here and here and the bug here.
So where does that leave us? I believe the best solution today is IFrames, which is what I went with. The article shahalpk linked is great but it only describes part of the process. Here's how I did it:
First, create an html file and js file for your isolated widget. Everything inside these files will run in an isolated environment in an iframe. Be sure to source your js file from the html file.
//iframe.js
var button = document.querySelector('.my-button');
button.addEventListener('click', function() {
// do useful things
});
//iframe.html
<style>
/* css */
</style>
<button class='my-button'>Hi there</button>
<script src='iframe.js'></script>
Next, inside your content script create an iframe element in javascript. You need to do it in javascript because you have to use chrome.extension.getURL in order to grab your iframe html file:
var iframe = document.createElement('iframe');
iframe.src = chrome.extension.getURL("iframe.html");
document.body.appendChild(iframe);
And that's it.
One thing to keep in mind: If you need to communicated between the iframe and the rest of the content script, you need to chrome.runtime.sendMessage() to the background page, and then chrome.tabs.sendMessage from the background page back to the tab. They can't communicate directly.
EDIT: I wrote a blog post detailing everything I learned through my process, including a complete example chrome extension and lots of links to different information:
https://apitman.com/3/#chrome-extension-content-script-stylesheet-isolation
In case my blog goes down, here's the sources to the original post:
Blog post
Example source

Either use all
.some-selector {
all: initial;
}
.some-selector * {
all: unset;
}
or use Shadow DOM
Library
function Widget(nodeName, appendTo){
this.outer = document.createElement(nodeName || 'DIV');
this.outer.className = 'extension-widget-' + chrome.runtime.id;
this.inner = this.outer.createShadowRoot();
(appendTo || document.body).appendChild(this.outer);
}
Widget.prototype.show = function(){
this.outer.style.display = 'block';
return this;
};
Widget.prototype.hide = function(){
this.outer.style.display = 'none';
return this;
};
Usage
var myWidget = new Widget();
myWidget.inner.innerHTML = '<h1>myWidget</h1>';
You can access the widget contents via myWidget.inner and the outer via myWidget.outer.
Styles
/*
* Reset Widget Wrapper Element
*/
.extension-widget-__MSG_##extension_id__ {
background: none;
border: none;
bottom: auto;
box-shadow: none;
color: black;
cursor: auto;
display: inline;
float: none;
font-family : "Helvetica Neue", "Helvetica", "Arial", sans-serif;
font-size: inherit;
font-style: normal;
font-variant: normal;
font-weight: normal;
height: auto;
left: auto;
letter-spacing: 0;
line-height: 100%;
margin: 0;
max-height: none;
max-width: none;
min-height: 0;
min-width: 0;
opacity: 1;
padding: 0;
position: static;
right: auto;
text-align: left;
text-decoration: none;
text-indent: 0;
text-shadow: none;
text-transform: none;
top: auto;
vertical-align: baseline;
white-space: normal;
width: auto;
z-index: 2147483648;
}
/*
* Add your own styles here
* but always prefix them with:
*
* .extension-widget-__MSG_##extension_id__
*
*/
.extension-widget-__MSG_##extension_id__{
position: fixed;
top: 100px;
margin: 0 auto;
left: 0;
right: 0;
width: 500px;
}
.extension-widget-__MSG_##extension_id__::shadow h1 {
display: block;
margin: 0 auto;
padding: 20px;
background-color: yellow;
border: 10px solid green;
font-size: 20px;
text-align: center;
}

I recently created Boundary, a CSS+JS library to solve problems just like this. Boundary creates elements that are completely separate from the existing webpage's CSS.
Take creating a dialog for example. After installing Boundary, you can do this in your content script
var dialog = Boundary.createBox("yourDialogID", "yourDialogClassName");
Boundary.loadBoxCSS("#yourDialogID", "style-for-elems-in-dialog.css");
Boundary.appendToBox(
"#yourDialogID",
"<button id='submit_button'>submit</button>"
);
Boundary.find("#submit_button").click(function() {
// some js after button is clicked.
});
Elements within #yourDialogID will not be affected by the existing webpage. And find() function returns a regular jQuery DOM element so you can do whatever you want with it.
Hope this helps. Please let me know if you have any question.
https://github.com/liviavinci/Boundary

Use iframes. It's a workaround, but works fine.
Maxime has written an article on it.

Related

Chrome Extension causing misbehavior with divs on active websites [duplicate]

I wrote a Google Chrome extension, which popups a dialog with an autocomplete field and it's own style, but there are some sites where my CSS gets totally broken, which doesn't look very nice.
I know about isolating styles with iFrames, but in Google Chrome extension there is no way to isolate my HTML and CSS in this way. Another method is to wrap all my stuff into a separated div with it's own id and relative styles for that id, and I do so, but it seems that it doesn't work on some sites with "hard" tags style overloading or "!important" directives in the CSS code.
So, I want to know is there any way to really isolate my styles in z convenient way or it's my bad carma to overload every little CSS property to fix one or another style issue for each site?
By the way: I set up my manifest to load all the things at the "document_end", but I see it's not being applied to the stylesheets which is every time loaded whenever the DOM is ready.
At the time of asking the question, your only option was to either use iframes, or stylesheets with a very high specificity and explicitly set all properties that might affect styles. The last method is very cumbersome, because there will always be some property that is overlooked by you. Consequently, the only usable method for isolating stylesheets was to use iframes.
The solution to this problem -isolation of styles without iframes- is Shadow DOM (since Chrome 25). You can find a tutorial at HTML5 Rocks. For a real-world Chrome extension that uses Shadow DOM to isolate styles, see Display #Anchors (source code here).
As I've recently gone through the gauntlet of this issue, I want to share some information I think is valuable.
First, Rob W's answer is correct. Shadow DOM is the correct solution to this problem. However, in my case not only did I need CSS isolation, I also needed JavaScript events. For example, what happens if the user clicks a button that lives within the isolated HTML? This gets really ugly with just Shadow DOM, but we have another Web Components technology, Custom Elements, to the rescue. Except that as of this writing there is a bug in chrome that prevents custom element in chrome extensions. See my questions here and here and the bug here.
So where does that leave us? I believe the best solution today is IFrames, which is what I went with. The article shahalpk linked is great but it only describes part of the process. Here's how I did it:
First, create an html file and js file for your isolated widget. Everything inside these files will run in an isolated environment in an iframe. Be sure to source your js file from the html file.
//iframe.js
var button = document.querySelector('.my-button');
button.addEventListener('click', function() {
// do useful things
});
//iframe.html
<style>
/* css */
</style>
<button class='my-button'>Hi there</button>
<script src='iframe.js'></script>
Next, inside your content script create an iframe element in javascript. You need to do it in javascript because you have to use chrome.extension.getURL in order to grab your iframe html file:
var iframe = document.createElement('iframe');
iframe.src = chrome.extension.getURL("iframe.html");
document.body.appendChild(iframe);
And that's it.
One thing to keep in mind: If you need to communicated between the iframe and the rest of the content script, you need to chrome.runtime.sendMessage() to the background page, and then chrome.tabs.sendMessage from the background page back to the tab. They can't communicate directly.
EDIT: I wrote a blog post detailing everything I learned through my process, including a complete example chrome extension and lots of links to different information:
https://apitman.com/3/#chrome-extension-content-script-stylesheet-isolation
In case my blog goes down, here's the sources to the original post:
Blog post
Example source
Either use all
.some-selector {
all: initial;
}
.some-selector * {
all: unset;
}
or use Shadow DOM
Library
function Widget(nodeName, appendTo){
this.outer = document.createElement(nodeName || 'DIV');
this.outer.className = 'extension-widget-' + chrome.runtime.id;
this.inner = this.outer.createShadowRoot();
(appendTo || document.body).appendChild(this.outer);
}
Widget.prototype.show = function(){
this.outer.style.display = 'block';
return this;
};
Widget.prototype.hide = function(){
this.outer.style.display = 'none';
return this;
};
Usage
var myWidget = new Widget();
myWidget.inner.innerHTML = '<h1>myWidget</h1>';
You can access the widget contents via myWidget.inner and the outer via myWidget.outer.
Styles
/*
* Reset Widget Wrapper Element
*/
.extension-widget-__MSG_##extension_id__ {
background: none;
border: none;
bottom: auto;
box-shadow: none;
color: black;
cursor: auto;
display: inline;
float: none;
font-family : "Helvetica Neue", "Helvetica", "Arial", sans-serif;
font-size: inherit;
font-style: normal;
font-variant: normal;
font-weight: normal;
height: auto;
left: auto;
letter-spacing: 0;
line-height: 100%;
margin: 0;
max-height: none;
max-width: none;
min-height: 0;
min-width: 0;
opacity: 1;
padding: 0;
position: static;
right: auto;
text-align: left;
text-decoration: none;
text-indent: 0;
text-shadow: none;
text-transform: none;
top: auto;
vertical-align: baseline;
white-space: normal;
width: auto;
z-index: 2147483648;
}
/*
* Add your own styles here
* but always prefix them with:
*
* .extension-widget-__MSG_##extension_id__
*
*/
.extension-widget-__MSG_##extension_id__{
position: fixed;
top: 100px;
margin: 0 auto;
left: 0;
right: 0;
width: 500px;
}
.extension-widget-__MSG_##extension_id__::shadow h1 {
display: block;
margin: 0 auto;
padding: 20px;
background-color: yellow;
border: 10px solid green;
font-size: 20px;
text-align: center;
}
I recently created Boundary, a CSS+JS library to solve problems just like this. Boundary creates elements that are completely separate from the existing webpage's CSS.
Take creating a dialog for example. After installing Boundary, you can do this in your content script
var dialog = Boundary.createBox("yourDialogID", "yourDialogClassName");
Boundary.loadBoxCSS("#yourDialogID", "style-for-elems-in-dialog.css");
Boundary.appendToBox(
"#yourDialogID",
"<button id='submit_button'>submit</button>"
);
Boundary.find("#submit_button").click(function() {
// some js after button is clicked.
});
Elements within #yourDialogID will not be affected by the existing webpage. And find() function returns a regular jQuery DOM element so you can do whatever you want with it.
Hope this helps. Please let me know if you have any question.
https://github.com/liviavinci/Boundary
Use iframes. It's a workaround, but works fine.
Maxime has written an article on it.

Custom styled checkbox (contact form 7) not working on iPhone

I've been working on a website that was built in WordPress using bootstrap templates. This website was not originally built by me, so I have troubles with it from time to time when implementing changes. The website uses yarn to build a 'dist' folder from node.js modules and elements contained within the theme folder. The stylesheets are scss files. After the build process finishes, it has created a 'dist' folder that is to be uploaded to the theme folder, which contains all the combined stylesheets, images, etc.
I'm having issues with a form that is created via contact form 7 and I'm not sure if this is an issue relating to the build process using yarn or something else. The issue is with a checkbox that I've added to the form. It works as expected on desktop displays, but when the site is viewed on an iPhone (via Chrome, Safari and Google browser apps) the checkbox tick doesn't appear on click.
Just for clarification, the last few websites I have built myself, that have included a form with a checkbox or checkboxes, are working perfectly across desktop and IOS devices. I use css stylesheets rather than scss, but I have used all the same styles on both the websites that are working fine and the one that is not working on IOS, so I really can't see why the checkbox is not working on mobile. I have also tried applying a couple of other fixes that have also not helped resolve the issue.
Please see below for the html code and scss code for the checkbox section of the website:
<div class="af-field af-field-type-checkbox af-field-checkbox acf-field acf-field-checkbox">
<div class="af-label acf-label">
<label for="brochure-pack-checkbox">Please send me a Giraffe Equity Release brochure pack</label>
</div>
<div class="af-input acf-input">
<div class="acf-input-wrap">
<span class="wpcf7-form-control-wrap brochure-pack"><span class="wpcf7-form-control wpcf7-checkbox" id="brochure-pack-checkbox"><span class="wpcf7-list-item first last"><label><input type="checkbox" name="brochure-pack[]" value="Please send me a Giraffe Equity Release brochure pack"><span class="wpcf7-list-item-label">Please send me a Giraffe Equity Release brochure pack</span></label></span></span></span>
</div>
<p></p></div>
<p></p></div>
.acf-field-checkbox {
.wpcf7-list-item {
position: relative;
margin-left: 0;
width: 100%;
display: block;
}
.wpcf7-list-item-label {
font-size: 2rem;
font-weight: 600;
line-height: 1.5em;
margin-left: 35px;
display: block;
cursor: pointer;
}
.wpcf7-list-item-label:before {
position: absolute;
left: 0;
width: 28px;
height: 28px;
border: 1px solid #b79c68;
border-radius: .25rem;
display: block;
}
.wpcf7-list-item-label:after {
position: absolute;
content: '\1F5F8';
font-size: 30px;
color: #b79c68;
width: 0px;
height: 0px;
top: 2px;
left: 2px;
opacity: 0;
}
input[type=checkbox] {
display: none;
-webkit-appearance: checkbox;
}
input[type=checkbox]:checked {
+ span.wpcf7-list-item-label:after {
opacity: 1;
}
}
label {
margin-bottom: 0;
cursor: pointer;
display: block;
}
}
I appreciate there may be some styles applied that are not needed, but they are not affecting the checkbox functionality. Once I have found the right solution to get this working for this specific website, I will clean up the styles applied and remove any that aren't necessary.
If anyone has any idea why the checkbox could be not working on this website, when they work fine on the others I have built, I would really be grateful for any hints or advice you could give.
Thanks in advance, if you need anymore information about the issue, then please ask and I'm sure I can clear up anything needed.
For anyone that is interested or has a similar issue, the cause of this problem was this:
.wpcf7-list-item-label:after {
content: '\1F5F8';
}
For some reason that code rendered a tick on desktop but not IOS. Changed to another tick code and resolved the issue.

Styled Components: Override Styles not working correctly with Server Side Rendering

It seems like overriding styles does not work consistently. I have this two styled components:
const StreamContentContainer = styled.div`
display: flex;
vertical-align: middle;
flex-direction: column;
flex: 0 100%;
align-items: center;
flex-wrap: wrap;
padding: 1rem;
flex-flow: column wrap;
`;
// new Component based on StreamContentContainer
with additional styles and override stlye
const FullStreamContentContainer = styled(StreamContentContainer)`
height: 56.11vw;
overflow: hidden;
padding: 0;
`;
Know if I use my FullStreamContentContainer there should be no padding. On CSR that works fine, the element shows up in the markup correctly.
But if the element is requested initial with SSR the padding of the StreamContentContainer overrides the FullStreamContentContainer again.
It does not matter if the node is generated on SSR or CSR, it shows up the same way in the markup:
<div class="sc-4y67w2-1 fodYop sc-4y67w2-0 WzHos">...</div>.
But if I inspect the element with the DevTools, I can see that on SSR first the class WzHos shows up and then the class fodYop:
Compared to the rules rendered on CSR the rules occur the other way round - like expected:
Does anybody know what causes this weird behavior and how to avoid it?
Problem resolved: Increased the specificity of FullStreamContentContainer styles by using
const FullStreamContentContainer = styled(StreamContentContainer)`
&&& {
height: 56.11vw;
overflow: hidden;
padding: 0;
}
`;
Relating to the SC docs, the repeated class bumps the specificity high enough to override the source order.

sharepoint branding : page width size due to _spBodyOnLoadWrapper() function

i'm trying to make up a different look and feel on my sharepoint site. I try to make my main content's width down to 960px under my form tag in sharepoitn designer. when I refresh the page at the first it renders the main content down to 960px but when the page fishished loading the main content stretches itself to the whole screen's width.
I found out that it's because of the onload script running in body tag. but I caanot remove this script because this work has side effects on page functionality.
the function is _spBodyOnLoadWrapper().
does anyone know this function ? or does anyone know how to come up with this problem ?
UPDATE #1:
My css code is as follows.. I added this class to the main Form on master page:
.mainContent
{
width: 960px;
height:100% !important;
min-height:100% !important;
padding-top: 10px;
margin-left: auto;
margin-right: auto;
direction:rtl;
}
UPDATE #2:
I use v4.master template.
I have taken the ribbon out of the form tag. it's directly after body tag. because I wanted the ribbon to be streched at the top. but when i add this line of code
<body onload="if (typeof(_spBodyOnLoadWrapper) != 'undefined') _spBodyOnLoadWrapper();">
the mainContent blows up. something at the header streches in the whole width and some panels at bottom remains 960px.
you got me interested so I recreated your issue. By default, the javascript will try to inline the width based on its calculations.
However, what you need is to set the class s4-notsetwidth on a wrapping container.
Here is what i did to fix your issue
Add this to the head
<style type="text/css">
#s4-bodyContainer {
width: 960px;
height:100% !important;
min-height:100% !important;
padding-top: 10px;
margin-left: auto;
margin-right: auto;
}
/* Aligns the Top Bars */
.ms-cui-ribbonTopBars {
width: 960px!important;
margin-left:auto;
margin-right:auto;
}
/* Turns off the border on the bottom of the tabs */
.ms-cui-ribbonTopBars > div {
border-bottom:1px solid transparent !important;
}
</style>
Then locate the s4-workspace (that's the immediate parent of #s4-bodyContainer and add class s4-nosetwidth. That should work for you.
Use these two references to achieve exactly what you want (not sure if you want ribbon aligned or not), Randy Drisgill post and Tom Wilson's post.

ckeditor 3 different body style for full page preview

Is it possible to change body style when entering a full page preview in ckeditor 3. Maybe to set a different body style for full page than when it is not in a full page mode.
The reason for this is using ckeditor when viewing a web page on a larger screen with maximized browser... in that case it is very wide and it is hard to read the content. So I would like to add in body style (but only for full page mode) something like:
...
margin: 5%;
padding: 5%;
border: 1px dotted #666;
...
... that will give more text processor look to the content.
TNX!
When CKEditor is toggled to fullscreen mode it adds "cke_maximized" class to container span.
So you may apply styles for entire container span (body+toolbar) like:
.cke_maximized{
margin: 5%;
padding: 5%;
border: 1px dotted #666;
}
or just for content body:
.cke_maximized iframe{
margin: 5%;
padding: 5%;
border: 1px dotted #666;
}
Those are just examples and you may experiment and choose css selector that is more suitable for you .
UPDATE 1:
Sure, you can use javascript code if it is not enough for your purposes. You can use something like this:
var editor = CKEDITOR.instances.editor1;
editor.on("afterCommandExec", function(e){
if(e.data.name == 'maximize'){
// maximized
if(e.data.command.state == CKEDITOR.TRISTATE_ON){
// add special css class to body(e.editor.document.getBody())
} else {
// minimized
// remove special css from body
}
}
});

Resources