Deciphering Col_flags4

[Robert The Rebuilder]

[quote="Nah'dneb"] I posted that last bit about warping a pinned gate object with collision because I had tested and verified it. Not because I was randomly throwing out Ideas. (like the rest of my posts in this thread)

[Nadnerb]

mass: 1
rc: 0
el: 0
col_type: 4
col_flags4: 00000004
alctype: collider
col_flags2: 02000000
col_flags0: 0200

all the other fields were either all zeros (colflags 1,3,5) or irellevant (name: RotWall)

oh, and a question. ptPhysics lists a "move" and a "rotate" function. have you managed to get either of those to do anything?

[Nadnerb]

due to your lack of response, I'm assuming that you are still having trouble with sliding colliders. now, I have tested such things in my own age and they worked perfectly. I decided to convert FuncChk01 to using a sliding collider. For some reason, It didn't work. after going over it many times, and more than a few gray hairs, I found that there there is a problem in here somewhere. The collider moved, but it was always in the wrong place. I think that it is a bug in the plugin. when I made a new region without the alctype:collider (visible) it worked. so then I changed it's alctype to collider and exported it again. It didn't work! but then I noticed something. (and I did this all over again to make sure) when I exported the age, it moved the ObData center of GateCollider. in blender! I can see it! WTF!? After multiple tries and much woe, I found the only thing to do was to change the python script to account for the displacement. So my conclusion (after testing) is that the plugin moves the object centers of -alctype:collider- objects to the center of the scene. Is this required by plasma? that all objects without a draw interface have center-of-the-scene centers? It doesn't do this to visible objects... and it still has a coordinate and simulation interface... (seeing as it does move and collide correctly) what's up?

before export
after export

[Robert The Rebuilder]

Nah'dneb:

Actually, I haven't had a moment to spend on that, although I really want to.

The reason behind the Python code that changes the center to be the origin for pinned colliders of type other than 3 is that Plasma will do a double transform on the matrix for the coordinate transform - thus displacing the object twice by its translation. However, when the center is placed at the origin, the matrices are the identity, and no double-displacement occurs.