kotlinlang #tornadofx

@amanda.hinchman-dominguez I fixed the felix bundle. I had an outdated version of configadmin, and they remove the old version from the repo when then upgrade. That broke the rest of the assembly as well.

edvin

11/06/2017, 1:00 PM

@pike I think you might just set the style right then and there, as the rowFactory callback will run quite often. In other words, just remove the listener and set the style based on the row item at that point in time.

edvin

11/06/2017, 1:00 PM

Your current approach will surely connect a lot of listeners to the same itemProperty.

pike

11/06/2017, 1:05 PM

Yea, but I need to search the results and highlight the row if it's found

amanda.hinchman-dominguez

11/06/2017, 1:28 PM

@amanda.hinchman-dominguez uploaded a file:

Screen Shot 2017-11-06 at 8.27.49 AM.png▾

amanda.hinchman-dominguez

11/06/2017, 1:30 PM

I should have posted a screenshot prior; orgSceneX/Y has no problem changing the value from initialization, but says in a reference (65 and 66) that it has no been initialized. I figured this had to do with some specific rules of type-safe coding

amanda.hinchman-dominguez

11/06/2017, 1:36 PM

I was modeling drag-and-drop shape events from this post http://www.java2s.com/Tutorials/Java/JavaFX_How_to/Shape/Handle_Shape_drag_and_drop_events.htm , and I was wondering what might be the best way to reference the type rectangle defined in tornadofx. I know rectangle defined in tornadofx transcribes as a Unit - perhaps I should try to define a type as any rectangle from Shapes.kt in the library?

carlw

11/06/2017, 1:37 PM

@amanda.hinchman-dominguez are you able to initialize sceneX and Y back on lines 54 and 55? I usually use a bogus value like -1

carlw

11/06/2017, 1:38 PM

if you want to defer the initialization, go with a nullable ? value

amanda.hinchman-dominguez

11/06/2017, 1:43 PM

If I assign a -1 to orgScene, that seems to solve the problem. Thanks! Should I try and create my own type to refer to a Unit rectangle?

carlw

11/06/2017, 1:45 PM

@amanda.hinchman-dominguez ~~you mean a standard blue 75x75 class ?~~ whoops...reading your link

edvin

11/06/2017, 1:46 PM

@pike You can still do that based on outside state, no need to attach listeners to the rowItem

edvin

11/06/2017, 1:48 PM

@amanda.hinchman-dominguez The compiler can't know that

mousePressed

will be called before

mouseDragged

🙂

amanda.hinchman-dominguez

11/06/2017, 1:48 PM

yeah, that's one I'm talking about - eventually, there will be a drag and drop between 2 panes (another challenge to tackle) but for now, I'll simply implementing the ability to drag and drop tornadofx-rendered shapes

carlw

11/06/2017, 1:49 PM

@amanda.hinchman-dominguez this is null-safety rather than type-safety (the initialization issue)

edvin

11/06/2017, 1:50 PM

It's actually not null-safety that's tripping the compiler. It's an uninitialized value, you'd have the same issue in Java 🙂

amanda.hinchman-dominguez

11/06/2017, 1:50 PM

the compiler can't know, because it doesn't matter what order it's called in?

edvin

11/06/2017, 1:50 PM

@amanda.hinchman-dominguez If you think about it, you could call those callbacks manually, so there is no guarantee about the order, and certainly not something the compiler can even know about 🙂

carlw

11/06/2017, 1:51 PM

@amanda.hinchman-dominguez it could infer an order, but something like "first defined" wouldn't be all that useful

amanda.hinchman-dominguez

11/06/2017, 1:51 PM

of course. Wouldn't these functions need to worry about what type of object they're grabbing though? I don't think the problem we're speaking of now can be solved until it knows what we're even grabbing

carlw

11/06/2017, 1:52 PM

@edvin i mean kotlin's null safety feature. java assigns unspecified values to null whereas kotlin doesn't

amanda.hinchman-dominguez

11/06/2017, 1:52 PM

I think these shapes fall under a unit-type, (which references components, right?) so I'd probably have to create a type based off shapes maybe?

edvin

11/06/2017, 1:53 PM

@carlw OK, maybe I misunderstood you 🙂 I meant to explain why the compiler didn't allow Amanda's code 🙂

amanda.hinchman-dominguez

11/06/2017, 1:53 PM

ahhh, @carlw now I get what you mean

carlw

11/06/2017, 1:53 PM

@edvin maybe i'm using the wrong term. in java "private Integer myint" is "private Integer myint = null". Kotlin doesn't do that

carlw

11/06/2017, 1:54 PM

@amanda.hinchman-dominguez you're trying to avoid the explicit cast to get at the underlying object?

amanda.hinchman-dominguez

11/06/2017, 1:55 PM

Not necessarily, I wasn't sure that I could cast for a shape - it was my thinking that since these shapes are Units (as are many other components), the compiler could be confused about what's a shape and what's not a shape