[fluxus] segfault when kicking a passive box

Dave Griffiths dave at pawfal.org
Fri Jul 14 07:18:05 PDT 2006


> Dave Griffiths wrote:
>>>On 14/07/06, Dave Griffiths <dave at pawfal.org> wrote:
>>>
>>>>you want to use fixed joints (build-fixedjoint object) which will
>>>>essentially lock the active object to the world - you can then connect
>>>>this to other active objects with normal joints.
>>>
>>>i though fixed joint still had to be between two active objects
>>>locking them together? and NB, ode documentation mention that using
>>>fixed joints is usually a bad idea (whatever they mean).
>>
>> creating fixed joints between two objects is a bad idea, I think just
>> because you should probably treat them as one object. fixed joints can
>> also be used in this other way though, to fix active objects to the
>> static
>> world. fluxus only exposes this way of using them, they were the only
>> way
>> I could find to "pin" things in a static way.
>
> Thanks for the tip, (build-fixedjoint obj) does exactly what I want it
> to do, once I figured out that the vectors for specifying locations of
> other joints don't respect the current transformations...

try (show-axis 1) to visualise the joints. I might change the joint space
to the current transform if it makes more sense. It certainly makes it
easier to procedurally grow physical objects that way, I've done it before
by manually transforming the position by (get-transform) in the script.

> Attached is my Newton's Cradle thingy, unfortunately the bounce is less
> than fulfilling... any ideas on how to make it less spongey?  I've tried
> fiddling with (surface-params) but to no avail, it's either more spongey
> or it goes completely crazy :/

it's always a trade off between stability and sponginess unfortunately,
that script is actually doing better than I'd imagine as it is. I might
steal this for the examples if you don't mind.

but yes - you generally have to push it to the edge of instability to get
the best result.

> Thanks for the Association List tip too, I think I'll find it very
> useful indeed :)

see also hash tables:
http://www.gnu.org/software/guile/docs/docs-1.8/guile-ref/Hash-Tables.html#Hash-Tables

cheers,

dave




More information about the Fluxus mailing list