KQL properties vs child nodes

[nichtich]

Given this KDL document

package name="foo" {
    name "bar"
}

What does KQL package name return?

• the value "foo"
• the value "bar"
• the values "foo" and "bar" (in this order)
• the values "bar" and "foo" (other way round)

Same question for package [] name (select package element, then "any element", then value of "name").

The result should be the same for package => prop(name) and package => name, right?

For package > bar the result should be the value "bar".

At least for package + name and package ~ name the result would be empty.

[zkat]

(note, we're working on new syntax, and >> is how we refer to descendants now. ++ is the same but for siblings)

Ok so.

KQL, with regular (non-map) syntax, only returns nodes. That is, package >> name and package > name will return the name "bar" node itself. Ditto for ANY selectors that select name.

The map operator, by the way, is going away with 2.0, but if you're implementing the 1.0 operator and want something like it, then yes, => prop(name) and => name are the same thing.

This is part of the reason I want to remove the map operator in 2.0: I really want people to be clear that what you're selecting is always nodes, not their properties or values. Think document.querySelector().

[nichtich]

Thanks and good to hear there will be a revised version of KQL. My next question would have been the kind of elements that KQL can return:

I really want people to be clear that what you're selecting is always nodes, not their properties or values. Think document.querySelector().

Or document.querySelectorAll(). With KQL you always select another KDL document (= a sequence of nodes) from a KDL document, don't you? Nice!