Re: [Pyode-user] dependence upon position?

[Matthias B. <ba...@ir...>]

mi...@mi... wrote:
> Strange. There is no 's' variable... apart from sx,sy,sz when setting glScale in the draw_body() part. Setting scale to zero would give an error, but they shouldn't be zero.

I also get such a message, but only for the version at the origin:

ODE Message 3: LCP internal error, s <= 0 (s=-5.2075e-003)

This is a message that originates from ODE (not PyODE), so the variable 
's' doesn't refer to any Python variable but rather to a variable that 
is most likely be used somewhere in the source code of ODE.

As one version shows the above error and the other doesn't, it is not 
very surprising that the results differ. Now as to why the error occurs 
I have no idea, I'd recommend to post a message into the ODE mailing list.

By the way, when I slightly rotate the cube (just so that it isn't 
perfectly aligned with the ground plane anymore) then the error does not 
occur anymore and the two simulations look the same.

- Matthias -

PS: I think the other case with the different start times is just the same.