Tab Atkins Jr.

The secondary solution (allow it to be escaped in parsing, but don't require it to be escaped in printing) appears to already be expected by the testsuite; see [`all_escapes.kdl`](https://github.com/kdl-org/kdl/blob/main/tests/test_cases/input/all_escapes.kdl) and...

oh lol I should look at commit histories; that was changed *by* the OP several days ago. :+1:

The alternate that was merged was just a change to the expected serialization for the tests; the larger (breaking change) question of whether to remove the `\/` escape entirely is...

I'm using entity now for the props/values object, and yeah, it works fine for me.

Yes, we could use a little more detail in precisely how the test suite expects numeric output to work; I had to do some slightly non-trivial stuff to hold onto...

In particular, right now the test suite expects: * numbers to be canonicalized to Decimal * the presence or absence of a decimal part and an exponent part to be...

> `+` isn't actually an issue, I don't think, since it's already not legal as the first char in a bare id. Yes it is, so long as the *second*...

> I don't think >> is too cumbersome, and as someone who doesn't do much complex css, it better communicates the meaning, imo. Fwiw, the CSSWG (aka, me, in this...

I'm fine with that. (We could keep `+` as an ident char but just remove it from *initial* ident chars if that would still be useful, so KQL could still...

Yeah, that sounds really good. It leaves KQL's evolution much more open without tying it so intimately to KDL's precise syntax model, and avoids restricting KDL itself unnecessarily. (Recall that...