Does Safari 15 finally fix viewport height?

Written by Luke Channings on June 11, 2021

TL;DR: No, but if you subtract env(safe-area-inset-bottom) from 100vh you'll get what you want.

Safari 15's UI is a radical departure from the previous version and from web browsers in general but does it fix the viewport height problem?

What is the viewport height problem again?

Mobile Safari has had problems related to 100vh not behaving like web developers expect1 2 pretty much since the beginning. The main crux of the issue is that Mobile Safari's UI Chrome shrinks when you scroll, and expands again when you activate it. That means 100vh ("100% the height of the viewport") can't be a static number.

Let's start by understanding the definition of the vh unit 3:

vh is defined as Equal to 1% of the height of the initial containing block.

Anthony Frehner

And here's the best explanation of the 100vh issues in Mobile Safari that I've seen so far,

The core issue is that mobile browsers (I’m looking at you, Chrome and Safari) have a “helpful” feature where the address bar is sometimes visible and sometimes hidden, changing the visible size of the viewport.

Rather than adjusting the height of 100vh to be the visible portion of the screen as the viewport height changes, these browsers instead have 100vh set to the height of the browser with the address bar hidden.

The result is that the bottom portion of the screen will be cut off when the address bar is visible, thus defeating the purpose of 100vh to begin with.

David Chanin, Avoid 100vh On Mobile Web

Let's put this new Safari to the test

I have a simple HTML page based on the example given in David's article. It has a header at the top and a button at the bottom, all wrapped in a 100vh container.

an image showing the aforementioned button hidden below the bottom UI controls in iOS 14's Mobile Safari
Safari 14 with Chrome active
an image showing the button being partially covered by the new floating address bar in iOS 15's Mobile Safari
Safari 15 with Chrome active

Safari's new floating address bar is displayed above our test button, which is more-or-less exactly the same behaviour as iOS 14.

So - Safari 15 does not change the behavour of 100vh 😢.

So what's the solution then?

It makes sense to me that the WebKit team wouldn't change the behaviour of the viewport unit, it's already well defined.

Do you remember when Apple introduced env() and safe-area-inset so that web developers could avoid their content being shown behind the notch4?

Well in Safari 14, safe-area-inset-bottom is 0px whether the UI chrome is active or inactive, which is something that has annoyed me for a while.

safe-area-inset-bottom is 0px when the UI chrome is inactive in Safari 15 on iOS, and then the height of the collapsed chrome minus the height of the expanded chrome when the bar is expanded.

That means that to get a button to float at the bottom of the page, always above the UI Chrome, all you have to do is use calc(100vh - env(safe-area-inset-bottom)).

Safari 14 with 100vh - safe-area-inset-bottom
Safari 15 with 100vh - safe-area-inset-bottom

Wrapping up

So not only does safe-area-inset-bottom work in Safari 15, it's animated!

I've been hoping that something to remedy the viewport height bug was coming since Jen Simmons (who joined the Safari / WebKit team in June 2020) was asking for feedback regarding viewport height issues.

It looks like this one's fixed!