NewtonCollisionCollideContinue missing edge collisions

[Julio Jerez]

it should compile fine

anyway, I think there is something wrong with google source, It keeps telling me that I do not have right to load my own file.

Synk to SVN until I can upload a version that you can just compile out of the box to google source.

[pHySiQuE]

I have always gotten the SDK from SVN. I am on revision 1530 now, the latest...

Working on it now...

[Julio Jerez]

I was making lot of changes to make the 32 and 64 bit build for both VS 2009 and 2010
I just download it for SVN and It build fine. all the 32 bit projects should be find. It is the 64 bit that had problems with some libraries in VS2010 but I just fixed them all

if you are syncing from SVN you you be at version 1531, try that and see if it works now.

[pHySiQuE]

When I include the .cpp files just like the Newton project does, and include the header file "coreLibrary_300/source/newton/Newton.h", this is my error log, with revision 1533:

log.rar

(11.27 KiB) Downloaded 165 times

[pHySiQuE]

Importing the precompiled static libs results in this:

Code: Select all

1>LIBCMTD.lib(invarg.obj) : error LNK2005: __invoke_watson already defined in MSVCRT.lib(MSVCR90.dll)
1>LIBCMTD.lib(dbghook.obj) : error LNK2005: __crt_debugger_hook already defined in MSVCRT.lib(MSVCR90.dll)
1>LIBCMTD.lib(setlocal.obj) : error LNK2005: __configthreadlocale already defined in MSVCRT.lib(MSVCR90.dll)
1>LIBCMTD.lib(setlocal.obj) : error LNK2005: _setlocale already defined in MSVCRT.lib(MSVCR90.dll)
1>LIBCMTD.lib(tidtable.obj) : error LNK2005: __encode_pointer already defined in MSVCRT.lib(MSVCR90.dll)
1>LIBCMTD.lib(tidtable.obj) : error LNK2005: __decode_pointer already defined in MSVCRT.lib(MSVCR90.dll)
1>LIBCMTD.lib(sprintf.obj) : error LNK2005: _sprintf already defined in MSVCRT.lib(MSVCR90.dll)
1>LIBCMTD.lib(winxfltr.obj) : error LNK2005: __XcptFilter already defined in MSVCRT.lib(MSVCR90.dll)
1>LIBCMTD.lib(dosmap.obj) : error LNK2005: __set_errno already defined in MSVCRT.lib(MSVCR90.dll)
1>LIBCMTD.lib(dosmap.obj) : error LNK2005: __errno already defined in MSVCRT.lib(MSVCR90.dll)
1>LIBCMTD.lib(crt0dat.obj) : error LNK2005: _exit already defined in MSVCRT.lib(MSVCR90.dll)
1>LIBCMTD.lib(crt0dat.obj) : error LNK2005: __exit already defined in MSVCRT.lib(MSVCR90.dll)
1>LIBCMTD.lib(crt0dat.obj) : error LNK2005: __cexit already defined in MSVCRT.lib(MSVCR90.dll)
1>LIBCMTD.lib(crt0dat.obj) : error LNK2005: __amsg_exit already defined in MSVCRT.lib(MSVCR90.dll)
1>LIBCMTD.lib(crt0dat.obj) : error LNK2005: __initterm_e already defined in MSVCRT.lib(MSVCR90.dll)
1>LIBCMTD.lib(strftime.obj) : error LNK2005: _strftime already defined in MSVCRT.lib(MSVCR90.dll)
1>LIBCMTD.lib(lconv.obj) : error LNK2005: _localeconv already defined in MSVCRT.lib(MSVCR90.dll)
1>LIBCMTD.lib(mlock.obj) : error LNK2005: __lock already defined in MSVCRT.lib(MSVCR90.dll)
1>LIBCMTD.lib(mlock.obj) : error LNK2005: __unlock already defined in MSVCRT.lib(MSVCR90.dll)
1>LIBCMTD.lib(printf.obj) : error LNK2005: _printf already defined in MSVCRT.lib(MSVCR90.dll)
1>LIBCMTD.lib(crt0init.obj) : error LNK2005: ___xi_a already defined in MSVCRT.lib(cinitexe.obj)
1>LIBCMTD.lib(crt0init.obj) : error LNK2005: ___xi_z already defined in MSVCRT.lib(cinitexe.obj)
1>LIBCMTD.lib(crt0init.obj) : error LNK2005: ___xc_a already defined in MSVCRT.lib(cinitexe.obj)
1>LIBCMTD.lib(crt0init.obj) : error LNK2005: ___xc_z already defined in MSVCRT.lib(cinitexe.obj)
1>LIBCMTD.lib(hooks.obj) : error LNK2005: "void __cdecl terminate(void)" (?terminate@@YAXXZ) already defined in MSVCRT.lib(MSVCR90.dll)
1>LIBCMTD.lib(crt0.obj) : error LNK2005: _mainCRTStartup already defined in MSVCRT.lib(crtexe.obj)
1>LIBCMTD.lib(errmode.obj) : error LNK2005: ___set_app_type already defined in MSVCRT.lib(MSVCR90.dll)
1>LIBCMTD.lib(_file.obj) : error LNK2005: ___iob_func already defined in MSVCRT.lib(MSVCR90.dll)
1>LIBCMTD.lib(getenv.obj) : error LNK2005: _getenv already defined in MSVCRT.lib(MSVCR90.dll)
1>LIBCMTD.lib(tolower.obj) : error LNK2005: _tolower already defined in MSVCRT.lib(MSVCR90.dll)
1>LIBCMTD.lib(fflush.obj) : error LNK2005: _fflush already defined in MSVCRT.lib(MSVCR90.dll)
1>MSVCRT.lib(MSVCR90.dll) : error LNK2005: __write already defined in LIBCMTD.lib(write.obj)

[pHySiQuE]

Ugh, it might help if I changed the 200 include directories to 300. The search path feature has to be the worst thing about C++.

[pHySiQuE]

Okay, initial convex casting results are pretty good. Nice work.

[Julio Jerez]

did you get it working?

[pHySiQuE]

Yes.

I perform raycasts on scaled meshes by creating a convex hull modifier for the sphere and scaling it, rather than re-creating a collision tree each time a mesh is scaled. I noticed the following:
-A convex hull modifier will not detect any collision if it is used instead of a sphere.
-If a plain sphere is used (instead of a convex hull modifier), it won't detect any collisions if the sphere radius is not equal on all axes.

Other than that, it seems very accurate and works well.

[Julio Jerez]

Yes that is correct, afte trying for almos 6 moteh tryon to figure out why an algorthm I found on the intenet did no work,
I decided to go back to my own solotion based of vonori reagion of the poligon.

My solution is not as fast by I am sure it is correct, teh Paul Netler seems very elegant abut I am afraid if has a flaw when extend to 3d and it simply fail.

anyway after, Ate I did some test I stop and move to complete some loose end I still have on teh physics part of teh engone.
you are correct an shere with diffenaxis size will fail ans so will be a modifirem teh reaso fo eth is that a Ellisos is a special case of a conve modiefer.

It is a very, eassy problem to solve, basically All I nee to do is to pretraform teh Face to teh space of the ellsion and made the call, on teh warp face.
I probably have teh call commnetd pout somewhere in teh code.
I will chek it out tonight.

also I am working of completting all of the Continue collision if the general engine because I need them for compound collision with small shapes (they get stock sometime),
and the Player controller which I starte to work and realize I did not port ove teh CC form core 200 yet.

I still have to add many of the sleeping optimization in core 300, as it is now the collision phase is doing brute force calculation of all pairs,
disrregading cache pair as ir did in core 200, but even doing that core 300 is much faster than core 200.

a feature that it occur to me be nice for core 300, is Uniform scale on collision shapes, there is no real reason not to suppost that nativally of the core shape. I will consider that.
A was resistance to sooprt that funtionality because of the sharing. but I
Very some I will add the coscept of Instancing to the collison shape. so we will be able to suooprt custom data on collison shape and still share information.
I actually do that for the I simple need to extend to the base class fo the shape.


core 200 had a problem with the broad phase supportion that, but core 300 does not, care 300 can fully arbitrary scale support along with free open spaces.

[pHySiQuE]

Sounds good.

core 200 had a problem with the broad phase supportion that, but core 300 does not, care 300 can fully arbitrary scale support along with free open spaces.

Does that mean 300 supports scaled matrices, or is this a feature you are going to add later?

[Julio Jerez]

NO, no scale matrices, the shape will suppost uniform scale withpout having to use modifier.
Maybe I remove the modofire all toghehet and all shape are modifiers, I still do no know what the lats solution will be, but Scale is a planed features.

another thing, on the continue collison for ellipsis and modifiers, do you think you cna make a similar test as you did before, that will help me a lot.

[pHySiQuE]

Yes, but I have a pretty big problem here with the 300 library:
viewtopic.php?f=12&t=7060

[pHySiQuE]

Problem fixed, I will see about getting you a demo now.

[pHySiQuE]

Okay, I sat down to make a demo for you, but the VS2008 project is giving linking errors when I build the debug DLL (300 library, revision 1549):

log.rar

(6.97 KiB) Downloaded 162 times