Этот сайт лучше всего просматривать в современном браузере с включённым JavaScript.

as3 and starling runtimes bounds incorrect

Erasmuz

As the title suggests I'm having difficulties getting bounds or measurements of a composed skeleton once it's built in starling.

Looking through the source for the runtime I see that the getBounds function was overridden, but incorrect values are being returned.
Also, getting the object's width or height always return 0 (assuming it might have something to do with the getBounds function, but didn't dig that deep).

The result of getBounds always sets the width and height to 0; no matter what (it's literally hard coded to return 0).

Is there some other way to get this data? Or is this just a bug? It looks like the bounds are calculated in the hittest function, so why not when the bounds are asked for?

I really need some way to capture the width and height; so please, is there some way to get access to this data?

EDIT: Just in case there's some other way to get the data I'm trying to get here's why I need the width and height.
1 - To calculate the center point of the image
2 - To accomplish scaling to get the image to fit in an exact space.
After some reading on the starling forum I read that you (Nate) did not calculate the width / height because it's expensive and the animation is changing, so the width and heights are also changing. Because of this hopefully there is something that is just currently eluding me to get the calculations I need. I can't be the only person who has ever needed to get this information.

Thanks.

Nate

Starling has a strange dependency where getBounds affects width, height, and the default hitTest implementation. You can read more here:
http://forum.starling-framework.org/top ... playobject
The way getBounds is implemented is necessary for rendering to work properly, even though it does not return the correct bounds of the current skeleton pose. To do that you'll need all the code from hitTest except the last IF, then use minX/Y and maxX/Y for your bounds.

Erasmuz

Thanks for the feedback!

That was exactly what I'm doing currently (extending your class and doing so instead of altering yours to prevent merge conflicts etc.)

I was just hoping I was missing something obvious to get these values.
Thanks again for the quick response, and awesome job on the software.

karpedm

I found that the overridden function "getBounds" in SkeletonSprite.as always returns zero width and zero height.
https://github.com/EsotericSoftware/spi ... te.as#L214

I also have some problem when I using starling filter on spine sprite. (spine object just disappears right after use filter)
And I found the reason is because of the wrong bounds value from SkeletonSprite which is essential for calculate the area when filtering.

I'm not sure if the problem is related with this.....

Nate

karpedm, you can see in the thread linked above that the Starling guys suggested the getBounds method as it is now. It is required that way for Starling to render correctly. I'm not sure if there is a workaround if it is causing problems with filters. You might have to ask the Starling guys.