Obi Official Forum

Full Version: ObiRope crashes
You're currently viewing a stripped down version of our content. View the full version with proper formatting.
Hello!

I get these crashes in random moments when I expend a rope. Could you help me to find out why? Sometimes it doesn't crash and sometimes does, but I can't find why.
Answer from support:
Quote:At first sight, the crash happens in a method called "Generate". The only method in our library with this name is not used for ropes, but for cloth, and only in editor. This is quite weird as rope should not be even remotely calling this method.
But the problem still exists.

I'm using version 3.2 of rope and 2017.1.0f3 version of Unity and I'm not using any cloth.
Logs are in archive here.
(17-10-2017, 07:05 AM)Evgenius Wrote: [ -> ]Hello!

I get these crashes in random moments when I expend a rope. Could you help me to find out why? Sometimes it doesn't crash and sometimes does, but I can't find why.
Answer from support:
But the problem still exists.

I'm using version 3.2 of rope and 2017.1.0f3 version of Unity and I'm not using any cloth.
Logs are in archive here.

Hi Evgenius,

Since the crash info provided by the log seemed bogus (as you state you're not using the function the log says is crashing in), we took a look at the dump files. All three say the library is crashing when trying to initialize a mutex, which is very weird. This kind of error (memory corruption) can come from invalid memory being fed to the library, or a memory alignment error that causes accesses outside of the intended memory region.

Can you give us more details about the crash? does it tend to happen when you do something specific? (you mentioned "expand" the rope, by pulling it or actually changing its lenght using ObiRopeCursor?)

Also, your platform seems to be a 32-bit one. Can you confirm this?
(17-10-2017, 04:01 PM)josemendez Wrote: [ -> ]Hi Evgenius,

Since the crash info provided by the log seemed bogus (as you state you're not using the function the log says is crashing in), we took a look at the dump files. All three say the library is crashing when trying to initialize a mutex, which is very weird. This kind of error (memory corruption) can come from invalid memory being fed to the library, or a memory alignment error that causes accesses outside of the intended memory region.

Can you give us more details about the crash? does it tend to happen when you do something specific? (you mentioned "expand" the rope, by pulling it or actually changing its lenght using ObiRopeCursor?)

Also, your platform seems to be a 32-bit one. Can you confirm this?

I have 4 ropes, 2 ropes attached with pin constraints to one object, to other 2 ropes attached to another object. I expand rope, detach the object that connected to those two (without changing rope connection), and then reel them up. Crash happens in random moment of that process. I'm changing length by using ObiRopeCursor. All ropes have the same length. Also all of them have one ObiSolver, if that could be a reason.

No, I'm using Windows 10 Pro, 64-bit.
(18-10-2017, 01:24 PM)Evgenius Wrote: [ -> ]I have 4 ropes, 2 ropes attached with pin constraints to one object, to other 2 ropes attached to another object. I expand rope, detach the object that connected to those two (without changing rope connection), and then reel them up. Crash happens in random moment of that process. I'm changing length by using ObiRopeCursor. All ropes have the same length. Also all of them have one ObiSolver, if that could be a reason.

No, I'm using Windows 10 Pro, 64-bit.

Hi there,

We've been trying to reproduce this on a variety of machines, ranging from Windows 7 to Windows 10, to no avail. We are using the "FreightLift" sample scene, as the setup is pretty much identical to the one you describe. We are doing this:

-Move the crane around
-Deactivate pin constraints, or remove them manually. Makes no difference removing 1,2,3 or all 4 of them.
-Move the crane around some more, lengthening and shortening the ropes.

At this point I think the best (if only) way to reproduce this is testing your scene. The logs you provided all crash at the same point, but it's a pretty weird spot (not a trivial fix, specially if we are unable to reproduce it). If you could share your project with us, that would be great.
The problem magically disappeared... I hope it's gone forever this time Sonrisa
(01-11-2017, 02:34 PM)Evgenius Wrote: [ -> ]The problem magically disappeared... I hope it's gone forever this time Sonrisa

Hi there! glad to know...another user had a very similar issue and surprisingly it seems it had to do with the nvidia GPU drivers. I don't know the exact cause but since both issues resolved spontaneously, I suppose both were related in some way.

Don't hesitate to tell us about it if it reappears.