Newton 2.0x Archemedia Open Beta

[Julio Jerez]

forget about 2.10 and 2.11 I was experimenting with VS10 new tools and there are lot of bugs,
I replaced with 2.12 please try that in was tested for days of consecuteve run and so far so good.

[bmwop]

why did the NewtonCreateTreeCollision build become twice or more slower between 2.09 and 2.10?

[Julio Jerez]

I unified soem aspect of teh class forn for collision tree an compound collision
now collision tree are mor eefficent at run time, bu teh nuild is a litle slower.

use 2.13 ther are more optimization

[JernejL]

why did the NewtonCreateTreeCollision build become twice or more slower between 2.09 and 2.10?

Use serialized trimeshes - it speeds up loading by factors of 1000x.

[bmwop]

Julio Jerez: 2.13 is the same as the 2.10 (from this point of view)

Delfi: wow, using serialize is quite good! Thanks!

[Julio Jerez]

2.10 is a particularly bad version full of Bugs, because I made lot of changes recomenbed by VS10

2.13 have bug fixes that are not on any previuos versions.

[Fold-away chair]

it might be a stupid question, and at the wrong location but:

with what do i run this library?
i cant seem to find any app or things.

[JernejL]

There are runable examples that come with the SDK.

[shybovycha]

2.10 is a particularly bad version full of Bugs, because I made lot of changes recomenbed by VS10

2.13 have bug fixes that are not on any previuos versions.

fixed bugs, which are not present in ANY version? =) lol

[JernejL]

It probably means that Julio added fixes that weren't fixed in any other version yet.

[shybovycha]

I'm sorry, i know that... I was just joking.... Sorry, again...

[JernejL]

No problem

[shybovycha]

BugReport: Ubuntu 9.10 x32, Newton 2.13, Crashes on Destruction demo... =(

[Sascha Willems]

After getting my TankDemo updated to the recent beta, I'm having troubles with the UserData in my material callback. Did anyhting change from the first 2.00 betas to the current ones on that part? I'm still doing it the same way :

Code: Select all

NewtonMaterialSetCollisionCallback(NewtonWorld, TrackLeft.MaterialID,  MaterialDefault, TrackLeft,  nil, @TrackContactProcess);
...
procedure TrackContactProcess(const material : PNewtonMaterial; const body0, body1 : PNewtonBody; pTimeStep : Float; pThreadIndex : Int); cdecl;
...Track := TTrack(NewtonMaterialGetMaterialPairUserData(Material));

But this isn't working anymore. Instead of getting the pointer to the track like it used to be I now get an increasing value each time the callback is run, which is very odd.

If I output it using this, result in a hex value that should be the same as the assigned userData :

Code: Select all

OutPutDebugString(PChar(IntToHex(Integer(TmpP),4)));

Instead I get increasing values that look like this :

Code: Select all

ebug-Ausgabe: 008A Prozess NewtonTankDemo.exe (1472)
Debug-Ausgabe: 008A Prozess NewtonTankDemo.exe (1472)
Debug-Ausgabe: 008B Prozess NewtonTankDemo.exe (1472)
Debug-Ausgabe: 008B Prozess NewtonTankDemo.exe (1472)
Debug-Ausgabe: 008C Prozess NewtonTankDemo.exe (1472)
Debug-Ausgabe: 008C Prozess NewtonTankDemo.exe (1472)
Debug-Ausgabe: 008D Prozess NewtonTankDemo.exe (1472)
Debug-Ausgabe: 008D Prozess NewtonTankDemo.exe (1472)
Debug-Ausgabe: 008E Prozess NewtonTankDemo.exe (1472)
Debug-Ausgabe: 008E Prozess NewtonTankDemo.exe (1472)
Debug-Ausgabe: 008F Prozess NewtonTankDemo.exe (1472)
Debug-Ausgabe: 008F Prozess NewtonTankDemo.exe (1472)
Debug-Ausgabe: 0090 Prozess NewtonTankDemo.exe (1472)
Debug-Ausgabe: 0090 Prozess NewtonTankDemo.exe (1472)
Debug-Ausgabe: 0091 Prozess NewtonTankDemo.exe (1472)
Debug-Ausgabe: 0091 Prozess NewtonTankDemo.exe (1472)
Debug-Ausgabe: 0092 Prozess NewtonTankDemo.exe (1472)
Debug-Ausgabe: 0092 Prozess NewtonTankDemo.exe (1472)
Debug-Ausgabe: 0093 Prozess NewtonTankDemo.exe (1472)
Debug-Ausgabe: 0093 Prozess NewtonTankDemo.exe (1472)
Debug-Ausgabe: 0094 Prozess NewtonTankDemo.exe (1472)
Debug-Ausgabe: 0094 Prozess NewtonTankDemo.exe (1472)
Debug-Ausgabe: 0095 Prozess NewtonTankDemo.exe (1472)
Debug-Ausgabe: 0095 Prozess NewtonTankDemo.exe (1472)
Debug-Ausgabe: 0096 Prozess NewtonTankDemo.exe (1472)
Debug-Ausgabe: 0096 Prozess NewtonTankDemo.exe (1472)
Debug-Ausgabe: 0097 Prozess NewtonTankDemo.exe (1472)
Debug-Ausgabe: 0097 Prozess NewtonTankDemo.exe (1472)
Debug-Ausgabe: 0098 Prozess NewtonTankDemo.exe (1472)
Debug-Ausgabe: 0098 Prozess NewtonTankDemo.exe (1472)

With none of the above values being the actual pointers to the assinged tracks. Since this worked fine some time ago, I'd like to know if something changed in terms of UserData or if there is a bug with that part of the current beta.

[JernejL]

Yep, the process callbacks changed greatly, see this - it should help you convert to new style callbacks & contact iterator properly: http://newtondynamics.com/wiki/index.ph ... 53_and_2.0