- From: Joe Pea via GitHub <sysbot+gh@w3.org>
- Date: Sat, 25 Jul 2020 01:58:26 +0000
- To: public-houdini-archive@w3.org
trusktr has just created a new issue for https://github.com/w3c/css-houdini-drafts: == [css-typed-om] == Continuing from https://github.com/w3c/css-houdini-drafts/issues/996, personally I was hoping that modifying `el.attributeStyleMap` would not be reflected back to the DOM `style` attribute, for performance gains. Seems half of that desire is already met in that `attributeChangedCallback` is not triggered. Can we prevent dynamic modification of the CSS Typed OM from modifying the DOM `style` attributes? Here's a test I made, [Bunnymark DOM V1 (CSS Typed OM via attributeStyleMap)](https://codepen.io/trusktr/pen/BajbvXr), in which I was hoping to reduce performance cost by updating continually-held references to the CSS Typed OM `CSSTransformValue`s of each bunny, then setting them with `bunny.attributeStyleMap.set('transform', transform)`. The more bunnies there are, the more DOM elements there are that will have updated `style` attributes. I'm not sure how much of a performance improvement not updating `style` attribute would have. At the moment, I see in Chrome's devtools Performance tab that most time is spent in `Recalculate Style` and `Update Layer Tree`. (I'm curious why those take so long.) Please view or discuss this issue at https://github.com/w3c/css-houdini-drafts/issues/997 using your GitHub account
Received on Saturday, 25 July 2020 01:58:29 UTC