PNS_PhysX [The only physX wrapper 4 linux] (QUAD/HEXO core)

devsh · Post by **devsh** » Wed Oct 14, 2009 6:18 pm

great xplanation....

btw, is there a way to apply brakes to a force in PhysX (i'm using them for half dynamic actors , i.e. my camera controller)

EDIT: nah character controller....

EDIT2: I think about doing contact report for character controller is it possible?

JP · Post by JP » Thu Oct 15, 2009 2:47 pm

I guess you'd apply brakes by just applying an opposing force...

and yes there's a seperate contact report system for character controllers, should all be explained in the character controller tutorials.

sio2 · Post by **sio2** » Thu Oct 15, 2009 7:37 pm

Halifax wrote:First off, I'd like to say it appears as though bitplane hit the nail right on the head.
BlindSide wrote:Another issue is of course the jp-sio2 incident which I would rather not get into.
I don't know if that was a joke...but there really is nothing to get into. sio2 simply released a modification of IrrPhysX branded as IrrPhysX without JP's approval. It could've been easily resolved, as JP had asked, by him submitting the changes to JP and allowing him to integrate them into his release schedule.

A misrepresentation of the facts. I made many posts stating what I was doing, leading up to a notification that I was ready to do a release - a post that JP himself replied to.

This is my last post on the matter. It's no problem to me at all to keep all my modifications from the last 10 months closed source.

Halifax · Post by **Halifax** » Thu Oct 15, 2009 8:02 pm

sio2 wrote:A misrepresentation of the facts. I made many posts stating what I was doing, leading up to a notification that I was ready to do a release - a post that JP himself replied to.

A release of code related to IrrPhysX is quite different when compared to a release of code that injects itself into IrrPhysX's release schedule as a new version. You even admitted yourself that you interpreted his reply wrong.

sio2 wrote: This is my last post on the matter. It's no problem to me at all to keep all my modifications from the last 10 months closed source.

sio2 wrote: I won't be making any more source releases or posts regarding IrrPhysx. I'll be continuing to use the library for private purposes as I already have a base implementation of character control, ragdoll and vehicle additions.

Really, it's nothing new.

Sorry, devsh, for the slight hijacking that I've caused. I'll get back on topic.

christianclavet · Post by **christianclavet** » Fri Oct 16, 2009 12:38 am

sio, why not release it? Just name it differently since you changed it a lot. SIOPHYSX wrapper for example...

devsh is this project can be easily ported to other platforms? Or you use specific system calls?

JP · Post by JP » Fri Oct 16, 2009 7:32 am

Or why not respond to the copious amounts of messages I've sent you so that we can get the changes put into IrrPhysx properly and released to the public... I thought you'd dropped off the face of the planet but now it's obvious you were ignoring me for some reason... Was there some unknown bust up that I wasn't part of??

devsh · Post by **devsh** » Wed Dec 02, 2009 9:34 pm

it turns out i needed to have some features and only then I could develop this wrapper...

PNS_PhysX has support for static triangle meshes with (UNLIKE irrPhysX) multiple mesh buffers, PNS_PhysX sticks them back together and passes that for cooking. Meaning that your large maps will be able to have static physics applied to them without the need of creating multiple physics actors (you would get a bug in irrPhysX and only the first meshbuffer/1000 vertices would have physics applied).

Tommorow will be Triggers and Raycasts. Which for example will allow me to implement a walking cycle into my player class. (knowing the slope etc.)

within this month we should have joints and in effect ragdolls.

sio2 · Post by **sio2** » Thu Dec 03, 2009 7:59 pm

JP wrote:Or why not respond to the copious amounts of messages I've sent you

I have two messages in my inbox from JP in relation to PhysX - one in Jan 09 and in May 09. TWO. Hardly copious.

Here's what JP posted in my thread AFTER I had posted an announcement in his thread stating that I was going to make a release and AFTER JP himself responded to my announcement "looking forward to seeing it":
"Right well i guess this shits all over IrrPhysx as done by me doesn't it? Gonna be a bit confusing if i release a 0.3 right? Would have been nice if you'd whispered a word in my ear before doing this..."

This really is the last I intend to comment on this. Apologies to devsh for the thread hijack...

JP · Post by JP » Fri Dec 04, 2009 9:55 am

To defend myself you never said you were going to actually release anything, I was expecting you to send it to me first so I could add it into IrrPhysx properly... simple as that... After the confusion was sorted out and it was clear there had been some misunderstanding I apologised for my outrage and asked for the source code which never got a response, and neither did any of my PMs or emails...

hybrid · Post by **hybrid** » Fri Dec 04, 2009 10:14 am

As I said in one of the other threads, you both should somehow solve this issue. Whether this means one or two projects is mostly irrelevant, although I'd prefer just one. MAybe simply set up a SF project or googlecode and work on the same code base. We could also put this project into irrExt, membership in that project is largely unrestricted.

devsh · Post by **devsh** » Mon Dec 07, 2009 9:41 pm

haha now a real feature irrphysX doesn't have

Visual Remote Debugging

yes you can download the app from nvidia and make my wrapper connect to it, without the pain of visualizing the data in irrlicht you can check if graphics match up with physX.

IT EVEN RUNS ON LINUX, yes the wrapper most certainly does but you can run the nvidia windows app in wine (which you cant do to physX

) and off you go debugging.

Personally I'm going to use it for my upcoming fearsome bones && joints (seriously it will be hell on earth) because it is very hard to have a physically correct mesh that behaves properly like a body when is hit by large flying objects, but doesnt fall on the ground like a dead body deprived of control.

My target is something like http://www.youtube.com/watch?v=_yGJPoTnK6c

P.S. Word of advice... (256x256Terrain && Linux && slow singlecore CPU && wine ) == VRD wine crash (your app doesn't so

for my design)