Stylis includes a prefixer module:
import { prefixer } from 'stylis';
This is very helpful as there are CSS properties that need vendor prefixes. For example, user-select: none; still requires it in browsers that have an important global usage rate.
However, I have noticed a couple of prefixes that I would argue are more harmful than beneficial. Mainly, the flexbox properties. The global need for the prefix seems to be 0.48% (0.01 + 0.14 + 0.05 + 0.01 + 0.02 + 0.03 + 0.01 + 0.1 + 0.11). Do we need to keep it, at this point?
By removing it, we would gain:
- A better DX when debugging CSS. When this module is used with emotion, you get this kind of output by default
Before

After

- A smaller bundle.
Actually, this might simply be the root cause of a different problem: What's the browser support target? Do we have something like this https://github.com/kripod/style-vendorizer#browser-support?
Stylis includes a prefixer module:
This is very helpful as there are CSS properties that need vendor prefixes. For example,
user-select: none;still requires it in browsers that have an important global usage rate.However, I have noticed a couple of prefixes that I would argue are more harmful than beneficial. Mainly, the flexbox properties. The global need for the prefix seems to be 0.48% (0.01 + 0.14 + 0.05 + 0.01 + 0.02 + 0.03 + 0.01 + 0.1 + 0.11). Do we need to keep it, at this point?
By removing it, we would gain:
Before
After
Actually, this might simply be the root cause of a different problem: What's the browser support target? Do we have something like this https://github.com/kripod/style-vendorizer#browser-support?