IPhysics - Newton/Irrlicht framework

TheRLG · Post by **TheRLG** » Mon Dec 04, 2006 7:53 pm

I very much liked this IPhysics. It seems very easy-to-use and probably fairly extendable. After getting it running with code::blocks and newton and everything, i was glad to compile successfully, but unfortunately everytime i ran it segfaulted at the part in the example code where you you created the actual physics node

i'm gonna have to play around with it more and see if i can get it running.

XFactor · Post by **XFactor** » Tue Dec 05, 2006 6:05 am

So I'm using Dev C++ because it seems to work with it. Is there a function for showing newton debug collisions. I remember someone had this before.

- Thank you

Nick_Japan · Post by **Nick_Japan** » Tue Dec 05, 2006 9:42 am

Hi all, glad this is still of interest. Sorry there have been no updates for a while - I've just started a new job and am just generally knackered at the end of the day. However, I have made some progress on character controllers and it well get done, slowly but surely.

About debug rendering, the Newton demos have that but it's part of their OpenGL drawing code, so I plan to use Irrlicht to do it. I'll try and make that part of the next release.

oldskoolPunk · Post by **oldskoolPunk** » Tue Dec 05, 2006 2:07 pm

Baiame wrote:Hmm, still doesn't work.

Irrlicht and Newton are set up properly, both work. In my project, IPhysics-1.2\lib is under "library directories", and IPhysics-1.2\include is under "include directories". I tried using the example 1 code and header, and every time I included IPhysics.h, I got a bunch of errors saying "In file included from...", a list, then "[Warning] no newline at end of file". The example fails to compile.

IPhysics.a is under the linker section on the parametrs tab of the project options. I was incorrectly using the .lib before, but I still get the same errors now.

Any help would be appreciated.

EDIT- I forgot to mention. All the errors returned are based on the IPhysics header files. I did try to add a new line to the end of all the headers, and it didn't work.

I am also getting this error, using the .a file. Was there ever a resolution?

veegun · Post by **veegun** » Tue Dec 05, 2006 5:07 pm

This is tight! I totally love this!

I had to rebuild the IPhysics.lib to target Newton SDK v1.53 and Irrlicht v1.2. That's because example 2 (terrain) kept dying on me during runtime.

Then I realize, hey, maybe IPhysics v1.2 is not updated to Irrlicht v1.2 or Newton v1.53? So, I rebuild the IPhysics lib and everything works now! Of course, I'm currently on a crappy PC that has no 3d support (intel integrated gfx) so I had to change from default OpenGL to software renderer.

What's with the wierd telescope view?

Major thumbs up, Nick! Maybe I should upload a video to youtube.

Nick_Japan · Post by **Nick_Japan** » Wed Dec 06, 2006 12:12 pm

Ah yes - sorry for all this confusion about the .a library. Basically, the one on my website is for IPhysics 1.1, so it won't work with the current release - I really should take it down. If you are using an IDE that doesn't play well with .lib libraries built with Visual C++, then for now either compile the IPhsyics source into a .a file yourself, or (easier in my opinion) just include the IPhysics source directly into your project and build from source. That disposes of the need for any linking - if you spend more than an hour trying to make it link, if I were you I'd really just do this instead. As IPhysics matures of course I will add libraries of all different creeds and kinds, but for now those are the options. Remember that IPhysics has no other dependancies; if you can compile an Irrlicht project and you can compile a Newton project, you can compile IPhysics.

I haven't tested IPhysics with Irrlicht 1.2, but I don't think there is anything that would break it - let me know if you have any trouble.

posttool · Post by **posttool** » Thu Dec 07, 2006 5:42 pm

Hi
You are really organizing the Netwon api in a very coherent manner. Thank you! !!!!

TheRLG · Post by **TheRLG** » Thu Dec 07, 2006 6:36 pm

Now, to only get more projects like this for PhysX and the like.. that would be nice.

RapchikProgrammer · Post by **RapchikProgrammer** » Wed Dec 13, 2006 8:04 pm

I love iphysics framework, and so i have already started working on it! I have my xams till 20th so progress will be slow from my end for now but i have already included materials! Nextup, ragdolls!

veegun · Post by **veegun** » Thu Dec 14, 2006 1:32 am

I wonder how hard it is to add other physics package to IPhysics? I would love to have IPhysics straight forward/clean interface for the other physics SDKs like bullet, ODE, and PhysX.

I know there's one of each already made by someone else on this forum. But a unified physics interface/API for irrlicht would be some much better.

I have my doubts of the other irrlicht physics authors would merge their respective physics with IPhysics.

Anyway, I was thinking of adding PhysX support to IPhysics that conforms to IPhysics current structure/layout.

Someone should also make an ISound that unifies all the different sound SDK/API out there (audiere, bass, irrKlang, FMOD, etc.).

Couple IPhysics and ISound with irrWizard and you have yourself a pretty nice package to make games with irrlicht.

RapchikProgrammer · Post by **RapchikProgrammer** » Thu Dec 14, 2006 6:44 am

To implement another physics library you would have to recreate iphysics from the base with abstract physics drivers classes! I was planning on making that kind of physics myself, but notsure when im gonna do that! Ill try to modify iphysics completely to make it easy for others to add other physics libraries in it too! But i hav vry little time nowadays so dont count on me!

monkeycracks · Post by **monkeycracks** » Thu Dec 14, 2006 9:58 pm

I tried recompiling IPhysics and I got this error.

In CPhysics.cpp :

Code: Select all

void * __cdecl physicsAlloc(int size)
{
    return malloc(size);
}

void __cdecl physicsFree(void *pointer, int size)
{
    free(pointer);
}

malloc undeclared
free undeclared

Anyone have any solutions? stdio.h IS included and everything is linked up and included properly...

Also, will it have a major impact on the way it works if I just comment it out? It compiles and runs if I do but I'm not sure if it will affect performance or not without those.

zeno60 · Post by **zeno60** » Thu Dec 14, 2006 10:09 pm

monkeycracks wrote:I tried recompiling IPhysics and I got this error.

malloc undeclared
free undeclared

Anyone have any solutions? stdio.h IS included and everything is linked up and included properly...

http://www.cplusplus.com/ref/cstdlib/

monkeycracks · Post by **monkeycracks** » Thu Dec 14, 2006 11:08 pm

zeno60 wrote:
monkeycracks wrote:I tried recompiling IPhysics and I got this error.

malloc undeclared
free undeclared

Anyone have any solutions? stdio.h IS included and everything is linked up and included properly...
http://www.cplusplus.com/ref/cstdlib/

Thanks zeno, that fixed it completely. I have the .a file for Irrlicht 1.2 and the latest Newton release if anyone wants/needs it, PM me. (Didn't see if this had been released anywhere else, if so then this will be redundant!)

veegun · Post by **veegun** » Thu Dec 14, 2006 11:10 pm

RapchikProgrammer wrote:To implement another physics library you would have to recreate iphysics from the base with abstract physics drivers classes! I was planning on making that kind of physics myself, but notsure when im gonna do that! Ill try to modify iphysics completely to make it easy for others to add other physics libraries in it too! But i hav vry little time nowadays so dont count on me!

So, how would you lay out the class structure if one was to add PhysX?

So instead of this:

// start iphysics
CPhysics physics;
physics.init();

We do something like this?

// start iphysics
CPhysX physics;
physics.init();

I was hoping that we could use the same class naming and do something like this:

// start iphysics
CPhysics physics;
physics.init(PhysX);

or

// start iphysics
CPhysics physics;
physics.init_physx();

What do you guys think?