[ODE] test_ode and capped cylinders

skjold@cistron.nl skjold at cistron.nl
Thu Feb 13 01:25:03 2003


Oh, it was about test_ode? I vaguely remember running it, and paying it no further attention because everything seemed to run fine (I just saw a lot of console output flash by, but at the time I was more interested in a visual demonstration of ODE). I'll run the test_ode tonight and chekc the output, just to be sure, but I suspect to be one of the lucky few that can run it without problems.

Greets,
Mark


> I have no input for you cylinder positioning problems, but I feel your pain
> with test_ode failing. I kept getting seg faults and other crashes on this
> program about a week ago. I mailed the list and the main response I got
> simply said something along the lines that the test_ode program is extremely
> sensitive and fails on a lot of machines. I was told as long as the
> box_stack demo works ok and doesn't have too much jitter and doesn't
> explode, then you're safe to use ODE. Apparently test_ode is meant as a test
> for Russ to judge how accurate the various ODE components are.
> 
> HTH
> John
> 
> -----Original Message-----
> From: ode-admin@q12.org [mailto:ode-admin@q12.org]On Behalf Of
> skjold@cistron.nl
> Sent: Thursday, February 13, 2003 2:30 AM
> To: ode@q12.org
> Subject: Re: [ODE] test_ode and capped cylinders
> 
> 
> Hi,
> 
> > Firstly, I'm having trouble placing the cylinders in my simulation. From
> > what I've read in the docs the second argument to
> > dMass::setCappedCylinder should dictate the cylinder's initial
> > directionm but no matter what I put there, the cylinder still starts out
> > vertically (along the z axis). Have I missed something, or do I have to
> > call a setRotation to get it the way I want?
> 
> The problem is, you can set an initial direction for the capped cylinder
> mass, but not for the capped cylinder geometry. I'm struggling with this
> oddity myself, but I think the only solution is to set the mass and geom
> vertically and then use dBodySetRotation to give it the orientation you
> want.
> 
> But I do feel compelled to add, that I would be a happy man if the
> dCreateCCylinder geometry function could be given that initial direction
> parameter as well, just like dMassSetCappedCylinder has.
> 
> 
> >
> > When trying to solve this, I noticed that my test_ode seg faults on all
> > the tests between textCholeskyFactorization() and testLDLTRemove(), and
> > the first test in testMassFunctions() fails. Is this a reason for concern?
> 
> I have no idea what you're talking about :P The only word I recognize here
> is seg fault, and it sounds bad ;) Sorry I can't be of help here.
> 
> Greets,
> Mark
> _______________________________________________
> ODE mailing list
> ODE@q12.org
> http://q12.org/mailman/listinfo/ode
>