Open kutensky opened 4 years ago
Drive by, but that's a really really expensive way to test for a pseudo-element. try { document.querySelector(":focus-visible"); return true; } catch { return false }
should be much faster.
I’m afraid this requires double your style declarations for .focus-visible
to make it work with or without the “polyfill” (technically this project isn’t a real polyfill).
I’m afraid this requires double your style declarations for
.focus-visible
to make it work with or without the “polyfill” (technically this project isn’t a real polyfill).
Yeap, that will require a double style declaration. But on the other hand, when browsers start to support his feature (Chrome is going to start supporting it from v.86), no js run will be needed. Currently, polyfill does a lot of background work that isn't good for performance.
Very good points from everyone.
It sounds like if we land this then we should use @emilio's implementation, but also land a note in the README which explains the need to double up your style declarations. We could also add a note that if folks would prefer, they can use the postcss plugin (https://github.com/csstools/postcss-focus-visible) which I think will let them write their CSS using :focus-visible
and then it'll do the right thing depending on their browser support matrix. I'm not a post-css expert but I think that's how it works :)
Related to #244
This condition could help
if (!window.CSS?.supports?.("selector(:focus-visible)")) {
// Apply the polyfill
}
Currently Chrome with experimental flag enabled supports ":focus-visible" polyfill. It would be nice if polyfill could be enabled only if current browser doesn' support ":focus-visible". To do that, I propose to check pseudo-class support using this method:
And then in polyfill modify code on the line 306 with:
After that we will be able to write css rules for both native ":focus-visible" and polyfill version: