Finally jME Physics 2.1 is released!

:smiley:



This release works with the jME 1.0 release and includes the ODE 0.9 binding, using ODEJava-JNI 0.3.1. The ODE implementation is the most complete binding for jME Physics. (JOODE and PhysX bindings are available from CVS)



Setup Tutorial



EDIT: updated link location

w00t! :slight_smile:

Any extra functionality such as limiting translation and rotational axii?

No. It’s pretty much like the release candidate. See readme.txt for a list of features.

PhysX as well… sweet… should have enough time to try it out next week or so…

I noticed on dev.java.net is says:


  • automatic generation of collision geometry (from boundings)



    Im guessing this applies to PhysicNode's function: generatePhysicsGeometry()? Is it possible to generate physics geometry from a model itself?
adamgp said:

Im guessing this applies to PhysicNode's function: generatePhysicsGeometry()?

yes


Is it possible to generate physics geometry from a model itself?

Well, you can use the trimesh itself and you can apply boundings and generate from those... what do you mean with 'from a model itself'?

Sorry I should have emphasised. I mean a detailed model which requires more accurate collision detection than using the models bounding box or sphere.

The idea was to have a detailed model for visuals and a low poly model which more or less maps the shape of the initial model for collisions.

And your question is… "Can jME Physics 2 generate a low poly physics model (probably optimized with some basic geometries) from a higher poly visual model as physical representation?"  Then the answer is: "No, you will need to decide what parts of your model are important yourself and provide that low poly optimized model."

Actually the question was; "I already have a low poly model, can I use this as the physics geometry for collisions?"

Of course you can use it! You did use jME Physics 2 already, didn't you :?

I cant download it even tho Im a member of dev,java.net :frowning:



I dont got "Project Page - View" premission :frowning:

Can someone give me that?

My username on the webbpage is Addez just as it is here



Please, I really need that physics…

It is now on googlecode:



http://code.google.com/p/jmephysics/



Tutorial:

http://www.jmonkeyengine.com/wiki/doku.php/setting_up_eclipse_to_build_jme_2#retrieving_jme_physics_2_and_correcting_build_path_problems



And one thing: If the obvious is not working (which is quite often the case) plz search the wiki and forum first.



@sbook/erlend: Could someone plz change the link of the sticky notes?

Addez said:

Please, I really need that physics..

ODEJava has a lot of bugs, if you really "need" physics, you will probably not get far with this. As soon as there are more than ten objects colliding you will get crashes..
I suggest trying my new jbullet-jme physics. Its alpha still, but normally does not crash like ODEJava does. See the link in my signature.
normen said:

...if you really "need" physics, you will probably not get far with this. As soon as there are more than ten objects colliding you will get crashes..

That is an inaccurate statement. But depending on needs, there may be more viable alternatives.
nymon said:

That is an inaccurate statement. But depending on needs, there may be more viable alternatives.

Well the number is certainly inaccurate, I am just judging from my own experience and that of other people, for example the Grappling Hook developer. I spent more than half a year with ODEJava and jmephysics2. The problem is that there is no defined problem or workaround for the sporadic crashes. The only workaround I could find personally was to have no collisions happening, which makes a physics engine pretty much useless. But since the problem is pretty obscure, there might be situations in which jmephysics2 works just fine without any crashes.

ttrocha, I use netbeans, and I cant find any tutorial on how to

install it on netbeans Sad



Is there any?