Skip to content

Instantly share code, notes, and snippets.

@paulirish
Last active March 19, 2024 07:51
Star You must be signed in to star a gist
Save paulirish/5d52fb081b3570c81e3a to your computer and use it in GitHub Desktop.
What forces layout/reflow. The comprehensive list.

What forces layout / reflow

All of the below properties or methods, when requested/called in JavaScript, will trigger the browser to synchronously calculate the style and layout*. This is also called reflow or layout thrashing, and is common performance bottleneck.

Generally, all APIs that synchronously provide layout metrics will trigger forced reflow / layout. Read on for additional cases and details.

Element APIs

Getting box metrics
  • elem.offsetLeft, elem.offsetTop, elem.offsetWidth, elem.offsetHeight, elem.offsetParent
  • elem.clientLeft, elem.clientTop, elem.clientWidth, elem.clientHeight
  • elem.getClientRects(), elem.getBoundingClientRect()
Scroll stuff
  • elem.scrollBy(), elem.scrollTo()
  • elem.scrollIntoView(), elem.scrollIntoViewIfNeeded()
  • elem.scrollWidth, elem.scrollHeight
  • elem.scrollLeft, elem.scrollTop also, setting them
Setting focus
Also…
  • elem.computedRole, elem.computedName
  • elem.innerText (source)

Getting window dimensions

  • window.scrollX, window.scrollY
  • window.innerHeight, window.innerWidth
  • window.visualViewport.height / width / offsetTop / offsetLeft (source)

document

  • document.scrollingElement only forces style
  • document.elementFromPoint

Forms: Setting selection + focus

  • inputElem.focus()
  • inputElem.select(), textareaElem.select()

Mouse events: Reading offset data

  • mouseEvt.layerX, mouseEvt.layerY, mouseEvt.offsetX, mouseEvt.offsetY (source)

Calling getComputedStyle()

window.getComputedStyle() will typically force style recalc.

window.getComputedStyle() will often force layout, as well.

Details of the conditions where gCS() forces layout

window.getComputedStyle() will force layout in one of 3 conditions:

  1. The element is in a shadow tree
  2. There are media queries (viewport-related ones). Specifically, one of the following: (source
    • min-width, min-height, max-width, max-height, width, height
    • aspect-ratio, min-aspect-ratio, max-aspect-ratio
    • device-pixel-ratio, resolution, orientation , min-device-pixel-ratio, max-device-pixel-ratio
  3. The property requested is one of the following: (source)
    • height, width
    • top, right, bottom, left
    • margin [-top, -right, -bottom, -left, or shorthand] only if the margin is fixed.
    • padding [-top, -right, -bottom, -left, or shorthand] only if the padding is fixed.
    • transform, transform-origin, perspective-origin
    • translate, rotate, scale
    • grid, grid-template, grid-template-columns, grid-template-rows
    • perspective-origin
    • These items were previously in the list but appear to not be any longer (as of Feb 2018): motion-path, motion-offset, motion-rotation, x, y, rx, ry

Getting Range dimensions

  • range.getClientRects(), range.getBoundingClientRect()

SVG

Quite a lot of properties/methods force, but I haven't made an exhaustive list. This list in incomplete:

  • SVGLocatable: computeCTM(), getBBox()
  • SVGTextContent: getCharNumAtPosition(), getComputedTextLength(), getEndPositionOfChar(), getExtentOfChar(), getNumberOfChars(), getRotationOfChar(), getStartPositionOfChar(), getSubStringLength(), selectSubString()
  • SVGUse: instanceRoot

Use the "chromium source tree link" below to explore on your own!

contenteditable

  • Lots & lots of stuff, …including copying an image to clipboard (source)

* Appendix

  • Reflow only has a cost if the document has changed and invalidated the style or layout. Typically, this is because the DOM was changed (classes modified, nodes added/removed, even adding a psuedo-class like :focus).
  • If layout is forced, style must be recalculated first. So forced layout triggers both operations. Their costs are very dependent on the content/situation, but typically both operations are similar in cost.
  • What should you do about all this? Well, the More on forced layout section below covers everything in more detail, but the short version is:
    1. for loops that force layout & change the DOM are the worst, avoid them.
    2. Use DevTools Performance Panel to see where this happens. You may be surprised to see how often your app code and library code hits this.
    3. Batch your writes & reads to the DOM (via FastDOM or a virtual DOM implementation). Read your metrics at the begininng of the frame (very very start of rAF, scroll handler, etc), when the numbers are still identical to the last time layout was done.

Timeline trace of The Guardian. Outbrain is forcing layout repeatedly, probably in a loop.
Cross-browser

More on forced layout


Updated slightly April 2020. Codesearch links and a few changes to relevant element properties.

@Losses
Copy link

Losses commented Apr 29, 2020

Another question, will these code trigger double reflow while user resizing their window?

window.addEventListener('resize', () => console.log(window.innerWidth))

One reflow may from window resize itself, another reflow comes from getting window.innerWidth

I wonder if the browser engine will merge these two reflow into one, this deeply confused me...

@EECOLOR
Copy link

EECOLOR commented Apr 29, 2020

@Losses The basic idea is that all of those methods will ensure they give the correct values. This means that if something has changed since the last calculation, they will calculate again to be sure.

A cycle in the browser consists of a few different sections (they go in circles, so me starting with one does not have any meaning) _ paint _ regular _ animation frame _ paint _ ....

With javascript we can execute code in 2 places: regular and animation frame. All code is not in requestAnimationFrame() will execute in regular. If you change something in the DOM and then ask the DOM about some measurement, it will calculate it for you. If you change something in the DOM before paint and ask the DOM about some measurement after paint it does not need to calculate, because it already knows.

So to be safe, make sure you only read from the DOM in regular and only write to the DOM in animation frame. Important: do not write to the DOM in regular and do not read from the DOM in animation frame.

This presentation helped for me: https://www.youtube.com/watch?v=cCOL7MC4Pl0

@Losses
Copy link

Losses commented Apr 30, 2020

@EECOLOR Wow, thanks for your answer, it really help!

@skortchmark9
Copy link

skortchmark9 commented May 5, 2020

Is there any way to avoid relayout due to changing text in a text node? I've tried many variations of fixed width/height, overflow, contain, monospaced fonts, all to no avail. I would even settle for being to limit the layout to the parent node, but right now it always seems to leak to the whole document...I'm literally considering using canvas for rendering text which just seems so wrong

@andfinally
Copy link

I'm interested to see matchMedia isn't in this list, since that must involve measuring window dimensions.

@melkishengue
Copy link

Great list, good to know !!

@htho
Copy link

htho commented Sep 8, 2020

I found another really nasty one: inserting <iframe> elements.

setTimeout(() => document.body.insertAdjacentHTML("beforeend", "<div><iframe src=\"\" style=\"display: none\"/></div>"), 2000);

@htho
Copy link

htho commented Sep 11, 2020

At least in jQuery 2.x the event handling component makes a copy of each property in an event (at least for click events). This includes the properties mentioned above.
This happens even though I have no jquery click events registered.
My workaround is to make a custom build of jquery, where I override the original copy routine with a routine that instead setups getters.

// someScriptBeforeJQueryIsIncluded.js

// activate workaround
window.FF_fix_useGetter = true;
// https://github.com/jquery/jquery/blob/2.2-stable/src/event.js#L445
fix: function( event ) {
	if ( event[ jQuery.expando ] ) {
		return event;
	}

	// Create a writable copy of the event object and normalize some properties
	var i, copy,
		type = event.type,
		originalEvent = event,
		fixHook = this.fixHooks[ type ];

	if ( !fixHook ) {
		this.fixHooks[ type ] = fixHook =
			rmouseEvent.test( type ) ? this.mouseHooks :
			rkeyEvent.test( type ) ? this.keyHooks :
			{};
	}
	copy = fixHook.props ? this.props.concat( fixHook.props ) : this.props;

	event = new jQuery.Event( originalEvent );

	i = copy.length;
	if(!window.FF_fix_useGetter) {
		// original approach
		var prop;
		while ( i-- ) {
			prop = copy[ i ];
			event[ prop ] = originalEvent[ prop ];
		}
	} else {
		// getter approach
		while ( i-- ) {
			const prop = copy[ i ];
			Object.defineProperty(event, prop, {
				get: function() { return originalEvent[ prop ]; }
			})
		}
	}

	// Support: Cordova 2.5 (WebKit) (#13255)
	// All events should have a target; Cordova deviceready doesn't
	if ( !event.target ) {
		event.target = document;
	}

	// Support: Safari 6.0+, Chrome<28
	// Target should not be a text node (#504, #13143)
	if ( event.target.nodeType === 3 ) {
		event.target = event.target.parentNode;
	}

	return fixHook.filter ? fixHook.filter( event, originalEvent ) : event;
},

@a-boertien
Copy link

Currently in Chromium: entering text input into a text area causes Layout:

https://bugs.chromium.org/p/chromium/issues/detail?id=1116001&q=textarea%20performance&can=2

@adrianoellero
Copy link

@aquaductape
Copy link

aquaductape commented Mar 2, 2021

I tested scrollY inside window scroll event on Chrome Version 88.0.4324.150 and it doesn't fire Layout calculation.

The page is very simple, has 2000 words, so that there's a scrollbar to test.
I ran two tests by using the Performance Tab. One that has window scroll event that logs scrollY and the other that doesn't have js files.

Here are the comparisons of Tasks.

The one that doesn't have js files.

Scroll > Event:scroll > Update Layer Tree > Composite Layers 

The one that does log scrollY on the scroll event.

Scroll > Event:scroll > Update Layer Tree > Composite Layers 
         Function call
         window.addEventListener.passive

The one that logs scrollY is almost identical, except for the js part where you see Function call. If there was Layout thrashing, inside the Task there would be 'Layout'.

Here's a demo https://8mxws.csb.app/
Here's the source https://codesandbox.io/s/testscrollyreflow-8mxws?file=/app.js

I didn't test within codesandbox, I tested locally, but i'm sharing it that way cuz it's the easiest way

@mfbx9da4
Copy link

mfbx9da4 commented Mar 29, 2021

If every item in the scrollable list has a &:hover {... will this cause layout thrashing when scrolling?

@lazysergey
Copy link

Scroll > Event:scroll > Update Layer Tree > Composite Layers
@aquaductape how do I get same output for debugging? any way to see what stages getting executed by browser?

@aquaductape
Copy link

aquaductape commented May 15, 2021

@lazysergey From what I know, the Performance result in chrome devtools doesn't format it in a text based way. The result is shown in GUI timeline and you hover over the cells and it tells what it is. So instead of showing a screenshot, I typed out the result as is.

But here's the screenshot in case

@lazysergey
Copy link

lazysergey commented May 17, 2021

thanks, for me they're also almost identical, I don't see any layout tasks

Screenshot 2021-05-17 at 8 44 02 AM

Screenshot 2021-05-17 at 8 41 54 AM

@waruyama
Copy link

waruyama commented Jul 1, 2021

Oh dear, it looks like changing the body's cursor with document.body.style.cursor = 'move'; invalidates the layout for the whole page. Ouch!

@nullhook
Copy link

nullhook commented Jul 1, 2021

@waruyama do you have a test link?

@waruyama
Copy link

waruyama commented Jul 2, 2021

@nullhook
Here is a little test case that creates an SVG with a massive amount of circles.

https://jsfiddle.net/17svhrqy/3/

  1. Run it and wait until "children created" is written to the console.
  2. Click on the red square and move the pointer.

If the cursor is changed on pointerdown, there is a gap of 500ms before mouse move event is triggered. Without changing the cursor (you have to comment that line), this gap is only 170ms. I am pretty sure that recalculation of the style is causing the difference.

@mootari
Copy link

mootari commented Aug 8, 2021

@htho
Copy link

htho commented Sep 6, 2021

I've started a little repository to reproduce these issues in order to learn from them. Maybe we can start to collect best practices there and see if some issues are resolved in the future.

https://github.com/htho/what-forces-layout-reflow

@matonga
Copy link

matonga commented Nov 25, 2021

Is there any way to avoid relayout due to changing text in a text node? I've tried many variations of fixed width/height, overflow, contain, monospaced fonts, all to no avail. I would even settle for being to limit the layout to the parent node, but right now it always seems to leak to the whole document...I'm literally considering using canvas for rendering text which just seems so wrong

@skortchmark9 thank you! I tried with an embedded SVG first, but it didn't do the trick. Using canvas was the only solution that worked. Anyway if someone knows some alternative I'd like to hear about it.

@dSalieri
Copy link

dSalieri commented Feb 11, 2022

Assignment to elem.style.transform value, invokes paint stage, also assignment to elem.style.left value, invokes paint stage twice. But if you will create property will-change, you don't see paint stage nowhere.

@Kaiido
Copy link

Kaiido commented Mar 11, 2022

A sneaky one I just found in the canvas2D API which currently only concerns Firefox:
setting ctx.fillStyle, ctx.strokeStyle or ctx.shadowColor will force a relayout there, when the associated canvas element is attached to the DOM.
This is because these values can accept a value of "currentColor", so the browser has to perform a recalc to get the computed color. Chrome and Safari have a bug where they only get that value from the canvas element's style attribute, so they don't need to trigger the recalc there. However their behavior is against the specs, so it may change in the future (or maybe the specs will change).

Test page: https://jsfiddle.net/sxe36Lk1/

@fabiospampinato
Copy link

fabiospampinato commented Apr 28, 2023

Weird how changing the focus triggers a style recalculation, like I don't even have any :focus or :focus-visible selectors, what's a style recalculation good for in this case? It looks like wasted time to me.

@Kaiido
Copy link

Kaiido commented Apr 28, 2023

@fabiospampinato it's the other way around. These getters and methods do force a reflow because they need an updated layout to work well, so it's not the fact that you have a :focus selector that matters1, but that the focus() method needs an up to date layout.
In the particular case of element.focus() for instance, it needs to check if the element is being-rendered, otherwise it's not a focusable-area, and thus can't be focused.

Footnotes

  1. Might be noted that the default UA stylesheet does have :focus-visible rule.

@fabiospampinato
Copy link

@Kaiido that makes a lot of sense 🤔 thank you!

@Fa-So
Copy link

Fa-So commented Apr 28, 2023 via email

@iamakulov
Copy link

The reason window.matchMedia is not on this list is because window.matchMedia("(min-width: ...)") does not depend on whether scrollbars are visible:

I believe that the list is correct and matchMedia does not trigger a reflow. All media queries have been carefully designed to not depend on page content. For example, width has been specced as viewport width including the scrollbar so that the result doesn't change regardless of whether there is a scrollbar or not.

@paulirish
Copy link
Author

A few notes on CSS Triggers

That site (which seems to be hosted by a squatter now?) covers what operations are required as a result of setting/changing a given CSS value. Basically, what work has to happen if you're changing just 1 style property every rAF tick.
"trigger" isn't a great name, but essentially it's poking at: what is invalidated (and thus needs to be recomputed) when style property X is changed?

For example:

  • setting el.style.opacity in a rAF loop doesn't invalidate the layout or paint, so on each rAF tick, we'd see Style then Composite. (skipping layout & paint)
  • meanwhile: setting el.style.left in a rAF loop affects layout as well. (And, as a rule, all layout changes need to be painted). So on each rAF tick we see all 4 operations: style,layout,paint,composite.

So while csstriggers is/was about "what is the impact of setting a style value?", the gist above is about "what is the impact of getting a value".

(tbh generally I think the csstriggers data is mostly a distraction/oversimplification and you should just use the devtools performance panel instead)

@Kaiido
Copy link

Kaiido commented Oct 27, 2023

A couple of new triggers I found, related to the Canvas2D API and my previous comment.

Another sneaky setter, that concerns both Firefox and Chrome: setting the Canvas2D .filter will force a reflow. https://jsfiddle.net/bqfm4agh/ I suppose this is because this property can also have a currentColor component (in drop-shadow()) and that they'll gather the computed styles every time.

Then, the Canvas2D API has a direction attribute, its default value is inherit, so it's taken from the computed style of the target canvas and thus everything that will get this attribute will need to trigger a reflow for connected <canvas> elements...

In Chrome the ctx.direction getter will trigger a reflow, that one is not very surprising. https://jsfiddle.net/mhn4y60c/

In Firefox the direction getter did not trigger a reflow, but that's because they computed it when we called getContext("2d") o_O https://jsfiddle.net/pdyc6v49/ That's certainly more surprising.

And then, in every browser, calling .fillText() or .strokeText() will also trigger a reflow. https://jsfiddle.net/80rgLymn/ You get it, they need the computed direction value for drawing the text, but while it makes sense when you know it, it's definitely not obvious (I discovered it very recently, while I'm interested on this kinda stuff for years now, and did work intensively on this API...).

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment