With the new Web Vitals incoming I have a problem with my sticky navigations/menus. In fact most pages will have :-(
The problem is that I use an approach like bootstrap affix to make the menu sticky when it would leave the viewport. But every time the menu leaves the viewport (and enters it as well) and the position is set from relative/static/absolute to fixed it increases CLS (Cumulative Layout Shift). I realize that changing the position to fixed will result in a layout shift because the element is removed from the layer and all following elements will be shifted to the top.
BUT: That's why I came up with some solutions and the best I think is that I use a wrapper with a specific height around the menu. So when the position of the menu changes to fixed the wrapper still exists and does not change in position or height, which means that no following elements has to shift. But the CLS is still counting up. And I do not know what to do to make my menus sticky without affecting the CLS which is important. Btw I cannot use position: sticky because there is not enough browser support. Because if my researches are correct position: sticky works without negatively affecting the CLS, my solution does not although the user **does not see any difference at all **....
Here comes some pseudo code to be more visual:
...
<body>
<h1>
Headline
</h1>
<p>
Here is some elements an stuff
</p>
<div class="menu-wrapper">
<div class="menu">
<ul>...........</ul>
</div>
</div>
<p>
More elements and stuff. Nothing shifting because the wrapper always has the same height.
</p>
...
</body>
...
.menu-wrapper {
height: 60px;
width: 100%;
}
.menu {
height: 60px;
width: 100%;
position: static;
}
.menu.affix {
position: fixed;
top: 0;
}
Any ideas? Thank you very much!
This may be fixed in Chrome Canary 90.
Context: CLS change log reports couple of improvements in this regards: https://chromium.googlesource.com/chromium/src/+/master/docs/speed/metrics_changelog/cls.md:
Cumulative Layout Shift Changelog
This is a list of changes to Cumulative Layout Shift.
Metric definition improvement: Bug fixes involving changes to transform, effect, clip or position
Metric definition improvement: Consider transform change countering layout shift.
Other fixes from Chrome 89 listed on the same page may apply also.
Related
Using bootstrap 3 I've got two columns (sm-3 and sm-9). The left column is position: fixed and contains a link with a tooltip. The z-index of the tooltip (without being specified) seems to be browser specific.
What is (in general) the best way to define the z-index of the tooltip? I'm looking for a clean solution that works in all common browsers.
FIDDLE
Internet Explorer 11 (left) Chrome 48 / Edge 25 (right)
HTML:
<div class="container">
<div class="col-sm-3" id="left">
<a id="tooltipButton" href="#" data-toggle="tooltip" title='Long Text'>Show Tooltip</a>
</div>
<div class="col-sm-9 col-sm-offset-3" id="right">
Content
</div>
</div>
CSS:
#left { position: fixed; }
#right { background-color: #ddd; }
JS:
$('#tooltipButton').tooltip({
trigger: 'click',
html: true,
placement: 'right',
});
A few things I've tried (CSS):
.tooltip { z-index: 10; } // Doesn't work
.tooltip .tooltip.in { z-index: 10; } // Doesn't work
.tooltip .tooltip-inner { z-index: 10; } // Doesn't work
Is there a clean solution without adding more divs and without changing the attributes of the left or right div? E.g. #left { z-index: 10; } seems to work but changes the attributes and the overlapping behavior of the left div.
FIDDLE
If you give an element a position other than static it forces it to appear above any element that is still static. It's a rather unknown behavior of z-index. Therefore if a parent element has a position other than static all it's children will have their own stacking scope. It works almost like layers.
This article has an amazing write up about it: What no-one told you about z-index
So in your example all the tooptip code is appearing in the container that has it's own stacking scope and is therefore forced under the next container. If you add z-index:999; to the #left container you will see it is fixed. This is because that container (and it's contents) are now forced higher than the next.
JSFIDDLE
The other thing you could do is remove position:relative from the #right container so it doesn't have it's own stacking scope and then it works again.
JSFIDDLE v2
I don't think this helps much with your problem as I don't think you can hack you way out of it. It is much more common practice that tooltip code is injected into the bottom of the page so that it naturally stacks above everything else.
When I shrink the width of my browser window (Firefox v26) so that only 1/2 of my home page is shown, the horizontal scrollbar appears on the bottom of the browser, which is fine.
But if I scroll the page to see its right half -- that right half is blank. In other words, after horizontally scrolling to the right (which moves the page's content leftward, obviously) -- the right side of the page does not redraw. It stays blank. Even if I hit the refresh on the browser URL bar.
I looked around and saw several posts. This one seemed to be exactly the same problem (only difference was, theirs involved the vertical scrollbar).
So I took the suggestion there -- which was to set my outermost content div (called wholePageDiv in the code below) and also my outerDiv to 'min-width: 100%" but this changed nothing.
Here's the very simple code:
<html>
<body>
<div id="wholePageDiv" class="wholePageDivForCentering">
<div id="outerDiv" style="margin: 0; margin-top: 10px; min-width: 100%;
display: inline-block; overflow: hidden">
(not shown: a bunch of divs with text)
</div>
</div>
</body>
<html>
Here is the wholePageDivForCentering CSS class, with the change made per that SO post I read:
.wholePageDivForCentering
{
/* width: 100%; */
min-width: 100%;
/* height: 100%; */
min-height: 100%;
white-space: nowrap;
margin: 0 auto;
}
I have looked at other websites to see if they exhibit the same "right side of scrolled page does not redraw" problem. On other websites I tested, I shrink the browser to 1/2 the width needed to show the whole page, then I scroll to see the right 1/2 of the page -- all other websites I check are successfully redrawing the right-side content as I scroll.
Do I have a CSS style problem above?
EDIT: I hit F12 in my browser and use the 'Inspector' tool and I clearly see that the only visible content is within the wholePageDiv and this div is not expanding at all, to the right, as I scroll to the right -- the Inspector shows that for whatever reason my outermost wholePageDiv is remaining the same fixed size as the viewport, and when I scroll to the right, this viewport outline as shown by the Inspector simply shifts leftward and does not expand on the right side to accommodate moving the scrollbar to the right.
I have added your html and css in a fiddle and it scrolls fine for me. The text of the div is displayed with no issues: http://jsfiddle.net/micahSan/UucLB/3/
same code as the OP
Can you replicate your problem in a fiddle so we can all see it?
I solved this (for now) by either hard-coding the div's width, or by programmatically increasing the div's width as the width of the browser window/document was changed. Hopefully will find a less kludgy solution later.
What is the best approach to restricting an absolutely positioned element's position, ideally in pure CSS?
I know that the following isn't possible but I guess what I'm looking for would look something like:
.stickyElement{
bottom-max:auto;
bottom-min:0px;
top-max: auto;
top-min: 100px;
}
That would allow an element to move to a position no less than 100px from the top of it's containing, positioned element.
An example (and my initial) use case for this is a menu that scrolls as part of a page but stops scrolling when it hits the top of a viewport. (Sticky menus?) an example of which can be seen on this page:
http://spektrummedia.com/startups
I fully expect that this is not possible without using some Javascript but I thought I'd put it out there.
position: sticky
There have been discussions in the W3C about this in recent years. One proposal is the addition of a sticky value for the position property.
.content {
position: -webkit-sticky;
position: -moz-sticky;
position: -ms-sticky;
position: -o-sticky;
position: sticky;
top: 10px;
}
This is currently supported in Chrome 23.0.1247.0 and later as an experimental feature. To enable it, enter about:flags for the URL address and press enter. Then search for "experimental WebKit features" and toggle it between enabled and disabled.
On the html5rocks website, there's a working demo.
Strictly speaking, this is an implementation of sticky content, and not a general-purpose way to limit the minimum or maximum position of an element relative to another element. However, sticky content might be the only practical application for the type of the behavior you're describing.
As there is no way to build this for all major browsers without the use of JavasScript I made my own solution with jQuery:
Assign position:relative to your sticky-top-menu. When it reaches the top of the browser window through scrolling the position is changed to positon:fixed.
Also give your sticky-top-menu top:0 to make sure that it sticks to the top of your browser window.
Here you find a working JSFiddle Example.
HTML
<header>I'm the Header</header>
<div class="sticky-top-menu">
<nav>
Page 1
Page 2
</nav>
</div>
<div class="content">
<p>Some content...</p>
</div>
jQuery
$(window).scroll(function () {
var headerTop = $("header").offset().top + $("header").outerHeight();
if ($(window).scrollTop() > headerTop) {
//when the header reaches the top of the window change position to fixed
$(".sticky-top-menu").css("position", "fixed");
} else {
//put position back to relative
$(".sticky-top-menu").css("position", "relative");
}
});
CSS
.sticky-top-menu {
position:relative;
top: 0px;
width: 100%;
}
I know this post is old, and I might be a little late to it, but to anyone still wondering how to do this i would suggest checking out the clamp() method in CSS, you could do something like this:
top: clamp(30px, 10vw, 50px);
Which would set the min top value to 30px, the ideal value to 10vw, and the max value to 50px.
A media query expression that defines the distance between body 0X 0Y and browser-window 0X 0Y would allow elements to be made sticky after page is scrolled
No such expression has otherwise been proposed and is not supported by any browser, to my knowledge, but it would be a useful expression to allow dynamic configuration of sticky elements, such as menu bars that are sticky after page is scrolled past head, without use of JavaScript.
.this element {
position: absolute;
top: 200px;
}
#media (max-scroll: 200px 0) {
.this.element {
position:fixed;
top: 0;
}
}
To my knowledge, there is no way to restrict an element that was positioned using absolute positioning using solely CSS.
I have two layout elements lets say one is 33%, the other 66%. They both use 100% of my screen size (So it is dependent on browser window). The smaller element also has a min-size property, so it cant fall below 250px;
Now if the layout is at least 757px large (so the size where the min property doesn't apply) everything looks fine. If the layout falls below the 757px the second element starts to overflow since it still takes the 66%, so the sum of both layouts exceeds the 100%.
I made some graphics to show the behavior:
Layout 1000px not overflowing:
Layout 500px overflowing
Is there a solution to prevent the overflow (not overflow: hidden) so the second element takes only the remaining space when the first element reaches it's min width.
Also JavaScript shouldn't be used excessive!
Regards, Stefan
Sure, this is actually pretty easy and requires a very minimal amount of code:
HTML:
<div class="sidebar">
...
</div>
<div class="content">
...
</div>
CSS:
.sidebar{
float: left;
width: 33%;
}
.content {
overflow: hidden; /* Ensures that your content will not overlap the sidebar */
}
You can view the live demo here: http://jsfiddle.net/7A4Tj/
Edit:
If you're trying to achieve a site layout that has equal-height background images for the sidebar and content, all you need to do is wrap both those elements in a containing div and use the Faux Columns technique.
Try using the following for the second widget:
position: fixed;
right: 0;
HereĀ“s my five cents
min-width on both divs
and a wrapper that also has min-width, plus both of the divs having percentage width
JS fiddle code
PS seems to be working fine in IE8
PPS Also do check out #media queries if you want to have conditional CSS rules on different window sizes, might be helpful. Will run on browsers supporting CSS3: CSS Media queries at CSS Tricks
Designer here, trying to code.
I am almost there! Trying to get a drop down menu from dynamic drive to work over an jQuery image rotator. Played with z-index. I can get the menu to work over the image rotation on all browsers except in IE compatibility mode, cannot click on the buttons in the rotator.
http://local495.bigrigmedia.com/
Any help would be greatly appreciated!
Always so much easier to get everything looking right in Photoshop eh? You can fix your overlap issue with 2 minor tweaks to the CSS:
styles.css
#top {
position: relative;
z-index: 2;
height: 155px;
}
ddsmoothmenu.css
.ddsmoothmenu{
position: relative;
z-index: 2;
/* remaining css */
}
homerotation.css
div#feature_list {
position: relative;
z-index: 1;
/* remaining css */
}
I also noticed you had a lot of z-index: -100 sprinkled around your CSS. Those are also going to cause you trouble. I would suggest taking them all out and just using the above 2 changes.
What the above 2 rules do is establish the stacking order for the menu and image rotator in a way that all browsers (including our friend IE) understands.
The trick with IE when using z-index is to make sure all the elements you're trying to overlap are in the same stacking context. IE creates a new stacking context whenever you use relative, absolute or fixed position on an element. In our case above, we're setting the stacking order on the top most elements in the stacking context (i.e. the document), therefore it will be respected.
Edit
Added a z-index to the #top container as this is actually the <div> that's at the same level in the document as <div id="feature_list">.