Building convex hull fails for a specific mesh (.obj included) and results in memory corruption about bepuphysics2 HOT 5 OPEN

should have a fix pushed tomorrow.

The ostensible fix was secretly possessed by satan; more time may be required.

from bepuphysics2.

Reproduced the asserts and some wonky behavior; should have a fix pushed tomorrow.

Notably, I couldn't reproduce the memory corruption. Not sure what's going on there, but there's a good chance the fix fixes that too.

from bepuphysics2.

Alright, should be fixed as of 0ecb874 (2.5.0-beta17).

The hulls created for that sort of pathological case should now be much less likely to contain spurious internal faces. I wasn't able to reproduce the infinite loop/memorystomp issue, but I did robustify the relevant loop's termination conditions. It no longer relies on any numerical assumptions and should rule out that particular failure.

Thanks for the report!

from bepuphysics2.

Thank you for the fix! I've backported this to our fork ( we're still on Mono under Unity ;_; But working on getting away from that ).

There seems to be the DebugStep debugging code left in through, we have noticed that it causes memory leaks when used. We have a PR for our fork that just removes it by commenting it out, but I'm not sure how you prefer to remove the debugging steps on your end?

Yellow-Dog-Man#3

from bepuphysics2.

Woooooooooops! I guess it was inevitable that I was going to forget to re-disable that one of these times. I'm at a conference, but I'll get that fixed as soon as I can; thanks for the report!

from bepuphysics2.