-
The @/supports rule is *intentionally* unaware of basically anything else. That has some limitations, like here, but there's a good reason for it. If support-checks need to *know* anything, then browsers have to maintain that knowledge, & checks can easily become dated or wrong. LeaVerou/1396142305527148552
-
So browsers intentionally keep the test simple: "can I parse this snippet of code, yes or no?" Any attempt to go farther — "is the code meaningful in context?" or "does it do what you want?" — would require building in all sorts of extra knowledge, making it brittle over time.
-
This is also why we can't test support for the
gap
property in aflex
context. What does "support" mean there? All browsers *allow* you to set gap with flexbox, or float on a grid item. They may not apply both at once, but it might be useful for fallbacks. That's supported!