[ODE] Triangle-box collider contribution

[asko]

Seems you're one step ahead of me :)

----- Original Message ----- 
From: "Jeffrey Smith" <jeffreys at Softimage.com>
To: "ODE Mailing List (E-mail)" <ode at q12.org>
Sent: Wednesday, February 25, 2004 10:00 PM
Subject: RE: [ODE] Triangle-box collider contribution


> I have fixed the problem John mentions, as well as added some of my own
code
> which combines potential contacts.  This results in many fewer contacts
> being passed out of dCollideBTL, as well as reducing or eliminating the
> asymmetry in some collisions (for example a trimesh box bouncing on a flat
> bbox surface).  Search for "GenerateContacts" to see what I've changed.
(I
> have left all the original code intact, merely ifdef'ing the appropriate
lines
> out).
>
> The file is available at:
>              http://www.smokingrobot.com/collision_trimesh_box.cpp
>
> -jeff
>
>
>
> -----Original Message-----
> From: ode-bounces at q12.org [mailto:ode-bounces at q12.org]On Behalf Of John
Miles
> Sent: February 25, 2004 1:38 PM
> To: asko; ode at q12.org
> Subject: RE: [ODE] Triangle-box collider contribution
>
>
> Thanks!  I'm extremely interested in any improvements to this code, and
will be happy to help test and debug any submissions.
>
> In this case, it looks like your code is trying to generate too many
contact points.  When I compile it into my ODE build and run either
test_moving_trimesh.exe or my own application, it tries to pass index values
to the SAFECONTACT macro that equal or exceed (Index & 0xffff) in line 916.
This condition fails the assert in line 114 of collision_trimesh_internal.h.
>
> It appears that you never check the low word of iFlags to determine the
maximum number of contacts to generate.
>
> -- jm
> -----Original Message-----
> From: ode-bounces at q12.org [mailto:ode-bounces at q12.org]On Behalf Of asko
> Sent: Wednesday, February 25, 2004 9:38 AM
> To: ode at q12.org
> Subject: [ODE] Triangle-box collider contribution
>
>
> Hello everyone,
>
> I've had the honour to port a piece of code, a mesh-box collider from the
as of yet unreleased Serious Engine 2 to ODE. From what I've experienced it
seems to be slightly more stabile, so it's worth a try if you have a lot of
box-triangle action going on :)
>
> I have uploaded the files to a zip:
> http://www.cad-creations.net/asko/TriBox/TriBox.zip
>
>
> Would be nice if someone with CVS access could add it to the contributions
section.
>
> _______________________________________________
> ODE mailing list
> ODE at q12.org
> http://q12.org/mailman/listinfo/ode
>