[PIXI] sprite.worldTransform[2] => undefined

[d13]

Hello,

 

I'm getting `undefined` with these values:

 

sprite.worldTransform[2]sprite.worldTransform[5]

According to this thread, they should return the DisplayObject's globalX and globalY positions:

 

http://www.html5gamedevs.com/topic/1962-getting-global-position-of-children/

 

Am I making a mistake somewhere?

 

[enpu]

I think that has changed. Try

sprite.worldTransform.txsprite.worldTransform.ty

[xerver]

 

I think that has changed. Try

sprite.worldTransform.txsprite.worldTransform.ty

 

Correct, we are using objects for worldTransform now not arrays.

[d13]

Thank you for your replies!

 

I tried sprite.worldTransform.tx and sprite.worldTransform.ty and they're both giving me the same value: 

 

0

 

If the sprite is at in a container at 10,10 and the container is on the stage at 50,50, shouldn't the worldTransform.tx and worldTransform.ty be 60,60?

Or am I misunderstanding these values?

(I'm using Pixi 1.5.3)

 

In my current project I'm calculating the stage positions of sprites in a container like this:

 

stageX = sprite.x + sprite.parent.x;

stageY = sprite.y + sprite.parent.y;

 

But this only works for one nested level. 

[enpu]

Is your container and sprite visible?

[d13]

Is your container and sprite visible?

Yes

[xerver]

Make sure that at least one run of `updateTransform` has gone through the entire scene graph. Remember this matrix is updated each frame render not when you change position/rotation values. That means you have to let a frame or two pass by then check the value.

[d13]

Make sure that at least one run of `updateTransform` has gone through the entire scene graph. Remember this matrix is updated each frame render not when you change position/rotation values. That means you have to let a frame or two pass by then check the value.

Thank you, I did not know this!

It works!!

[d13]

... but now that it works, I have a related question

Using the `worldTransform` values I sometimes see off-by-one inaccuracies in my collision detection functions.

(Using the parent x and y offsets is fine, and so are local x and y if the sprites are in the same coordinate space.)

 

Could this have something do with the order of evaluation in `updateTransform`? 

[xerver]

... but now that it works, I have a related question

Using the `worldTransform` values I sometimes see off-by-one inaccuracies in my collision detection functions.

(Using the parent x and y offsets is fine, and so are local x and y if the sprites are in the same coordinate space.)

 

Could this have something do with the order of evaluation in `updateTransform`? 

 

Off-by-one as in if you add the tree yourself you get a different value than what is in the transform matrix?

[d13]

Off-by-one as in if you add the tree yourself you get a different value than what is in the transform matrix?

 

I think I know what's happening.

As you explained, `updateTransform` updates `worldTransform` each frame, not at the instant when a sprite's local x/y position is changed.

That means you can't accurately use `worldTransform.tx` and `worldTransform.tx` for collision detection and reaction because they contain the position values from the previous frame.

I am understanding this correctly?

[xerver]

I think I know what's happening.

As you explained, `updateTransform` updates `worldTransform` each frame, not at the instant when a sprite's local x/y position is changed.

That means you can't accurately use `worldTransform.tx` and `worldTransform.tx` for collision detection and reaction because they contain the position values from the previous frame.

I am understanding this correctly?

 

That is correct.

[d13]

Great, now I know how this is all working.

I'll see if I can devise some other way of calculating immediate worldX and worldY co-ordinates that isn't too expensive.

... maybe by peering up through the parent/child hierarchy... I'll test that