@tieskedh I added an overloaded launch function that accepts

Array<String>

as well:

fun main(args: Array<String>) = launch<MyApp>(args)

No need to use the spread operator 🙂

I also relaxed it to accept javafx.application.Application subclasses.

great. I'm now looking into the streams. This needs to be checked, before accepting!!! Afterwards, is it an idea to use the onchange we provide to the outside throughout the framework (where possible), instead of calling the native setonchange?

@tieskedh There is a small bug in PR 432, I started a review and added a comment on it.

@tieskedh I've done that in some places I think so I guess it makes sense, but since it actually results in a deeper call stack I'm not sure it's worth it. Some times it's worth doing extra work in the framework itself. On the other hand, people might look to the framework code for best practices. The "performance hit" is negligible, so I say go for it 🙂

@edvin on this moment the framework only consists out of onChange with one param. I like to add onChange with multiple params, but Kotlin has not decided to stick to their own statement about it. (I'm clearly opinionated about what they should do 😉 ). See also https://discuss.kotlinlang.org/t/overload-ambiguity-lambda-parametercount My question is, should we wait and hope that kotlin quickly makes up it's mind, should we warn users not to use it in combination with onChange and after some iterations add onChange (so we can be consistent in the code) Or should we add functions with other names, to add these listeners right away?

I'm heading to San Francisco this week to shoot a Kotlin for Data Science video. Here is my TornadoFX example, which will be the grand finale and the final example for the course:

@User uploaded a file: Untitled

@User uploaded a file: Untitled

@User uploaded a file:

@User uploaded a file:

@thomasnield Nice example! Not much code for that UI 🙂

inline fun <reified T : View> EventTarget.add(type: KClass<T>, params: Map<*, Any?>? = null): Unit = plusAssign(find(type, scope, params).root)

what does reified do here?

A little research for issue 340 https://courses.bekwam.net/public_tutorials/bkcourse_greatgrid_fx_and_swing.html

@tieskedh I think it's to support the find() method which has to use the T::class meta information in one of its forms to provide a class arg.

@tieskedh This snippet shows inline and reified in action. If you take off the reified, you won't have access to the T::class info (it won't compile). In TornadoFX, it's not reified everywhere because some of the calls involve non-inline functions and kotlin doesn't support reified across-the-board.

@User uploaded a file: Reified Demo

yes, but in the example, it's not used I believe? Or is it also coupled to the type?

btw, plusassign is not reified

This is the find i'm referring to

inline fun <reified T : Component> find(params: Map<*, Any?>? = null, noinline op: (T.() -> Unit)? = null): T = find(T::class, scope, params).apply { op?.invoke(this) }

it's a convenience that turns around and calls another non-reified find()

I get the basic reified and try to add it (don't know if it will be accepted) my example is https://github.com/edvin/tornadofx/blob/master/src/main/java/tornadofx/Component.kt on line 680 this code looks odd, because in my eyes it doesn't use the reified T. The reason why I'm asking is that there was earlier odd--looking code, with a very good reason, so I try to understand if this is just odd code, or if there are things I don't understand yet.

On second thought, I just pulled it off the functions in my copy of Component.kt. I don't think they're needed in EventTarget.add(). Those calls have a class object.

I was able to compile without error

I think you're right, thx

@tieskedh @edvin I went through the Git commits in Component.kt. The EventTarget add methods were added in 198261b5e309271a089d8f261ef02b548aca8f21 on 10/24/2016. It looks like we've always passed a type in to all versions of TornadoFX, so I think they can be safely removed.

@tieskedh I think you're right, reified does nothing here, even inline could be removed. Maybe this function used to call find with type parameter instead of passing the class.

brb