Should we have child.textMaybe <-- ... ?

[yurique]

or, maybe, child.text.maybe ?

[raquo]

How would you guys feel about a single child <-- method that accepts (observable: Observable[A: Renderable])? where Renderable[A] is basically A => ChildNode?

I'm not a fan of conceptual complexity (implicits and stuff) but it might be worth it here. That would also work similarly outside of child <--, with an implicit def renderableToMod[V: Renderable](v: V): Mod[ReactiveElement.Base].

[yurique]

I like it.

Maybe this will also help scalac figure out stuff like this:

div(
  Option("a").getOrElse("b")
)

[raquo]

I'm now wondering if there's a good reason to keep child.text separate from child at all. Due to our implicits, we can already say child <-- fooObservable.map(_.toString) and that will work because an implicit conversion will be inserted into _.toString. But child <-- stringObservable will not work because there's no place to inject the implicit. Which feels unintuitive to someone not familiar with our implicits.

Our child.text already has def <--[T]($node: Source[T])(implicit ev: T => TextNode), we might as well just make child <-- itself into def <--[T]($node: Source[T])(implicit ev: T => Child), then we won't need child.text at all.