Help with Reliable Crash in Rigid Body Scene

33 views
Skip to first unread message

Matthias Worch

unread,
Jan 6, 2008, 3:26:25 AM1/6/08
to x...@softimage.com
I have a scene with rigid bodies that crashes reliably when saving. Which obviously sucks badly. Very simple setup, just a passive body that serves as the ground and a bunch of falling sheets that land on it. And a bunch of forces in the scene.

Worked fine for a while, now every time I try to save the scene the progress bar goes to 100% and then crashes with "Successfully saved the scene before system failure." The resulting scene file is either empty or corrupted. There's also a log warning at the bottom of the screen: "Unable to resolve path: T:\dsInfineonTurn07IRL\Simulation\RigidBodyCache.xsi". The fun thing is that this is *not* my current project, whatever XSI is searching for is in a different project folder. So that might be a hint as to what is going on.

Does anybody have an idea what to do about this? I haven't done much Dynamics in XSI.
I'm on v6.02 and can't really upgrade to 6.5 right now - but any idea if this was fixed?

todd akita

unread,
Jan 6, 2008, 3:49:59 AM1/6/08
to X...@softimage.com
Matthias Worch wrote:
> There's also a log warning at the bottom of the screen: "Unable to
> resolve path: T:\dsInfineonTurn07IRL\Simulation\RigidBodyCache.xsi".
> The fun thing is that this is *not* my current project, whatever XSI
> is searching for is in a different project folder.
You can change that path in the 'external files' menu. It's also
usually not a good idea to go deleting cache pointers, no matter how
unnecessary they may seem.

-T

---
Unsubscribe? Mail Majo...@Softimage.COM with the following text in body:
unsubscribe xsi

Matthias Worch

unread,
Jan 6, 2008, 4:01:20 AM1/6/08
to X...@softimage.com
I've been playing around with those options just now, but no matter what I do (internal cache, external .xsi, loaded/offloaded etc) the scene still crashes at the end of the save. It also crashes if I delete the Scene_Root_Environment_Cache for the scene - which might or might not be expected (and AFAIK is what XSI does automatically at the end of the save operation).

Hmm. I can rebuild the simulation easily enough in a new scene, would actually have been faster than trying to troubleshoot this. I'd just like to know how I got XSI into this state in the first place so that I can avoid it in the future...

kim aldis

unread,
Jan 6, 2008, 4:08:59 AM1/6/08
to X...@softimage.com

It may just be twisted in the scene somewhere and a rebuild may not be a total disaster. Have you tried either merging the scene into a new one or saving the sim into a model and importing it back in. If the latter works it may be the scene rather than the sim and its objects. Maybe also try deleting the sim objects and saving as a further check.

Matthias Worch

unread,
Jan 6, 2008, 4:16:42 AM1/6/08
to X...@softimage.com
I've tracked this down to one specific rigid body. XSI crashes if I try to delete the mesh from the scene, XSI crashes when I try to remove the rigid body from that mesh. Merging the scene with a new one doesn't get rid of it, either - sounds like a test case to send to Softimage.

kim aldis

unread,
Jan 6, 2008, 5:44:30 AM1/6/08
to X...@softimage.com

It’s worth sending off but it looks like it may be a corruption of some kind.

peter boeykens

unread,
Jan 6, 2008, 9:23:29 AM1/6/08
to X...@softimage.com
i've seen this kind of behaviour in totally different contexts,
usually with a couple of objects in a modelling relation- like one object generated from another (eg. extract polygons), and then a third related to the extracted one (eg with shrink wrap.)
when you organise those into nested models, you can end up with corrupted models.
I've seen some where every action, including selecting it in the explorer without any viewport open, was enough to make XSI disappear instantly.
My solution was exporting everything else from the scene as a model and getting that into a new scene, rebuilding the offending object from that.
 
So yes about the exporting as a model - this should become almost second nature for anything even of low complexity.
Reply all
Reply to author
Forward
0 new messages