On Thu, Aug 21, 2008 at 12:30 AM, Kao Cardoso Felix <[EMAIL PROTECTED]> wrote:
>
> On Tue, Aug 19, 2008 at 10:11 PM, Kao Cardoso Felix <[EMAIL PROTECTED]> wrote:
>> On Tue, Aug 19, 2008 at 10:02 PM, Lucio Torre <[EMAIL PROTECTED]> wrote:
>>> As you are using your own images, you have to set the image anchor
>>> yourself at the center. when you create sprites using:
>>>
>>> sprite = Sprite("myimage.png")
>
> Ok, so I read the code in cocos/sprite.py and when I pass an image
> this is what cocos do:
>
> if isinstance(image, str):
> image = pyglet.resource.image(image)
>
> well, it's exactly what I did in my code and after that the image
> anchor is never touched again until this:
>
> if anchor is None:
> if isinstance(self.image, pyglet.image.Animation):
> anchor = (image.frames[0].image.width / 2,
> image.frames[0].image.height / 2)
> else:
> anchor = image.width / 2, image.height / 2
>
> wich sets my anchor independent of I loading the image myself or
> passing a string to cocos.Sprite(). So it really doesn't matter if I
> load the image myself.
>
> Well, I read a good part of the cocos source to understand what was
> going on on my simple program. Well I found the root of the strange
> behaviour at the CocosNode transform method code. I made some
> simplifications (ommited those optimization tests and camera
> transform) just fore the sake of this discussion and the code does
> something like this:
>
>
> glTranslatef(
> self.position[0] + self.transform_anchor_x,
> self.position[1] + self.transform_anchor_y,
> 0
> )
>
> glScalef( self.scale, self.scale, 1)
> glRotatef( -self.rotation, 0, 0, 1)
>
> glTranslatef(
> self.children_anchor_x - self.transform_anchor_x,
> self.children_anchor_y - self.transform_anchor_y,
> 0
> )
>
> This is moving the child node by the vector that points from
> transform_anchor to child_anchor. Then it's doing the rotation and
> scaling on the node. Last, it moves the child by the offset of the
> parent position plus the transform_anchor.
>
> Well the thing is that my sprite's transform_anchor and
> children_anchor coincide (because they are set by the anchor property)
> and are in the middle of the sprite. In the last transformation I only
> revert the transform_anchor offset applied before and not the
> children_anchor offset, thus my sprite is moved exactly by the
> positive value of children_anchor wich, in my case, is my sprite
> center, so it ends up in the top right corner of the parent sprite.
>
> Please, note that I described the transformations in their order of
> ocurrence wich would be from the bottom to the top since we are
> dealing with OpenGL.
>
> Well, I found out that including the negative children anchor offset
> on the last transformation corrects the strange behaviour I have seem
> and, aparently, don't break any of the demos. I written the code like
> this:
>
> glTranslatef(
> self.position[0] - (self.children_anchor_x -
> self.transform_anchor_x),
> self.position[1] - (self.children_anchor_y -
> self.transform_anchor_y),
> 0
> )
>
> glScalef( self.scale, self.scale, 1)
> glRotatef( -self.rotation, 0, 0, 1)
>
> glTranslatef(
> self.children_anchor_x - self.transform_anchor_x,
> self.children_anchor_y - self.transform_anchor_y,
> 0
> )
>
> Of course the optimization tests are still missing, but I found it
> much easier to reason about the code this way.
>
> BUT, there is a problem with my solution above. The children_anchor
> property isn't working exactly right. For instance, when I change only
> the children_anchor of my sprite to (0,0) and mantain the
> transform_anchor in the middle (w/2, h/2) the child still is on the
> center of the parent, but rotations happens around a point that is
> offset by the difference between the children_anchor and
> transform_anchor. Based on the names of the two properties it seems
> that their roles are swaped.
>
> I hope I made myself clear enough. I prepared a minimalistic test code
> that causes the bug. The only thing I couldn't provide with it are the
> two images I used to test, but anyone would to the trick. The code is
> here:
>
> http://pastebin.com/m25cd6671
>
> Many thanks for your attention and sorry for this very long email.
>
hmmm. ill need some more time to look at this.
regards,
lucio.
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"cocos2d discuss" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to [EMAIL PROTECTED]
For more options, visit this group at
http://groups.google.com/group/cocos-discuss?hl=en
-~----------~----~----~----~------~----~------~--~---
