I'm working on a CSS library that includes hundreds of selectors and rules for quick templating instead of writing regular CSS. This is used as a replacement of attr()
CSS function until it works for any rule and not just content:
.
At first I generated a minified CSS file to load. But as there are many loops for each possible values, I've been trying to generate those CSS rules directly in the browser instead.
Here is a sample of a performance result:
- Prebuilt CSS file : 150KB / 15ms script CPU usage
- JS file (injecting in
<style>
element) : 20KB / 150ms script CPU usage
Of course those results can vary a lot depending on the network and the device. But in overall, what would be the best choice between network optimisation and CPU optimisation (for any kind of device)?
Aucun commentaire:
Enregistrer un commentaire