What does this error mean?

[FSA]

Oh: It's recommend to klick "Fenstermodus" (Windowed)

[Julio Jerez]

I am debuggind this now, and when attach the debuger both release and debug I do no get those error.

you sa teh initial location is 1, 1, 1

but I get (-25, -0.0029046535, -.05

+ veloc {...} dVector
step 0.00000000 float
+ veloc {...} dVector
+ matrix {m_front={...} m_up={...} m_right={...} ...} dMatrix
- (*(TemplateVector<float>*)(&(matrix).m_posit)) {m_x=-25.000008 m_y=-0.0029046535 m_z=-0.50909072 ...} TemplateVector<float>
m_x -25.000008 float
m_y -0.0029046535 float
m_z -0.50909072 float
m_w 1.0000000 float


are you sure is this the demo that produce the bug? because I wolk around planty I I do not get that problem.
I made a smal chnmage to teh code so that is supports arbitray playe origin. Plase get later SDK and try again.

edit.
Oh I see what you mean, when I run form teh cpmmand lien teh bug does happens but, if I run form visual studio is does not.
very strange.

OK I still debugging this.

[FSA]

Yes that's what i mean. From VisualStudio an debug build it works. From Explorer/coommandline not. The intial Position is not 1|1|1 but i get these coordinates after the start. Tank you for Debugging

[Julio Jerez]

I do nor undernat why this happen, I have never seen a bug like this.

If is something the teh convex cast, if I use the debug library, I do get an assert, but only form the command line, if I step in debug I do no get stack or any debug info
if I run form veusil studio teh bug does no happen, I will see if I can do it with remote debugging .

I am guessing that some how ther is a function that is overwriting the return adress of a function.

[Julio Jerez]

I do not know what this is, it does not crash in I run in remote debugging. bu the same program crashes from teh command line, and when I attach ther si no stak infromation.
tell me something is you program a c charp application that has debug option disallowed?

[Julio Jerez]

Yp have to have a linking bug in in you setting, it is not just that it crashe there, in craskk in any funtion call if I insesrt and assert .

also when I manage to step in teh joint library dl. and IO step in any of teh Newton DLL I get waring that teh souce code that tah wa suse to buidl teh library is no the same.
usually that means tha some is wrong, in the lib table of function.

can I see the include source code that you are using to set this, because I do no see how to debugg it.
bascially all pointe and argumen in amost all funtion are off by few bytes.
this is wha the velocity is wrog,

[FSA]

Sorry I did not understand this. You mean that i use old source but new libaries? I will update everything wait...

[FSA]

Must i link core.lib and others? Or just dMath, dJointLibary, newton?

[Julio Jerez]

No visual stutio tells me that the suorce that was use to compile the library is different that the source tha is in my system.
I says that if I debug the joint library and I try to step in the NewtonDLL. or if I debud the newton DLL and try to step in the Joint Libarry.
somehow there is a mimatch in the link table in release mode.

try to synk to SVN and link again, it may be that I made change to the Jopint library and you link your exe to an old .lib file and this is why the parametes are out of phase.
I also make sure that all library has consistent compile setting, but that was not the problem.
if the Bug is still there, I think there only way to debug it is by getting the source of your test program.
Because the crash happen even when using Remote debugging,

I have never seem a bug that even remote debug fail to catch, something is very bad.

[Julio Jerez]

Must i link core.lib and others? Or just dMath, dJointLibary, newton?

since you are not compiling teh code, you just nee to make sure the .lib tah you are using are the correct one and that they are upto date

[FSA]

I recompiled newton and the other projects in release dll. Overwritten the libaries and headers, recompiled my Engine, recompiled the "test". Same result as before.
New link: https://www.dropbox.com/s/eaea65gjkk14w ... io_new.zip

[FSA]

Doesnt read the post from you wait I will make a svn checkout...

[FSA]

Here the new link with everything recompiled and svn checkout: https://www.dropbox.com/s/c9fqaybzdccfw ... ly_new.zip

[Julio Jerez]

the link seem wrong.

[FSA]

Oh no Here the new one: https://www.dropbox.com/s/hcgjslimgb0km ... _Julio.zip