[ODE] non-capped cylinder support (dcyl2) now available in UNSTABLE

[Jaroslav Sinecky]

Hi!
I was observing one problem in Cylinder - Trimesh collision. If you drop a
short cylinder with larger radius (like pneumatics shape for example) on
triangle with cylinder axis having something between 20 and 70 degrees with
triangle normal, it first misses collision, later finds it when cylinder is
already more penetrated and gets pumped up, resulting in the cylinder
jumping around in quite unnatural way.

I found the bug (dFabs instead of dSqrt) and put my suggestion of how to fix
it on SF:
http://sourceforge.net/tracker/index.php?func=detail&aid=1354935&group_id=24
884&atid=382801


Apart, I've been using dCyl2 for couple weeks and it's very nice!

Cheers,
Jaroslav



> -----Original Message-----
> From: ode-bounces at q12.org [mailto:ode-bounces at q12.org]On Behalf
> Of Bram Stolk
> Sent: Thursday, November 17, 2005 3:47 PM
> To: ode at q12.org
> Subject: [ODE] non-capped cylinder support (dcyl2) now available
> in UNSTABLE
>
>
> Hi,
>
>
> The UNSTABLE branch now has dcyl2 incorporated.
>
> dcyl2 was a contrib, (by Alen Ladavac, ported to ODE by Nguyen Binh),
> that is now in the regular source tree of UNSTABLE.
>
> I've added a test_cyl.cpp test app that tests cylinders against
> trimesh world, and a box.
>
> To use non-capped cylinders, do the following:
>
> ./configure --with-opcode --with-cylinder
>
> I've been careful not to break anything, so a configure run
> without the flag should not be affected at all.
>
> This has only been tested on linux.
>
>
> Please report your findings,
> Later on, I plan to remove the contrib/ entry from UNSTABLE.
>
>    Bram Stolk
>
> PS: single-precision ODE seems to have an issue with test_cyl
>      sometimes collisions are not detected.
>      I'm pretty sure this is a trimesh issue, and not a cyl issue.
> _______________________________________________
> ODE mailing list
> ODE at q12.org
> http://q12.org/mailman/listinfo/ode