I’ve been trying to package my game for 2 days and I keep getting “Unknown Cook Failure” errors.
I’ve Ctrl+F’d my way through the log and, while warnings abound, the only time the word “ERROR” ever actually appears is in a string of identical lines in sequence, multiple times in the log:
LogPhysics:Error: PHYSX: (d:\ori.cohen_g6217_physics\engine\source\thirdparty\physx\physx-3.3\source\physx\src\NpRigidBodyTemplate.h 437) eINVALID_PARAMETER : RigidBody::setRigidBodyFlag: kinematic bodies with CCD enabled are not supported! CCD will be ignored.
Cook log attached:
[Cook Log.txt][1]
Now, I’ve gone through and checked everything I can think of. I can find no static meshes with CCD enabled by default, no Physics Assets with CCD on ANY of their bone bodies, no blueprints which use CCD whatsoever, no triggers or blocking volumes using it, etc.
I’ve also gone through just-about everything that’s changed since the LAST time I cooked looking for compile errors or warnings, and found nothing. Fixed up all redirectors, nothing.
SO, obviously, part one of this is to say “what else can I do to find out what object is throwing these errors so I can uncheck CCD on it?”
Part two is to ask, why on earth was this changed in 4.12? I’ve never run into this problem cooking before with respect to CCD being enabled on kinematic bodies. The log even SAYS “will be ignored”, which to me seems like the appropriate behavior, rather than killing the cook process with an unknown error. It seems like a pretty silly reason to throw a build-destroying error since it produces no problems when playing standalone.
Part three is to ask, why on earth doesn’t the log tell me which assets are causing this issue? Near as I can figure (by walking up dozens and dozens of lines) is that the problem occurs when cooking a level. That level has probably THOUSANDS of unique elements in terms of actors, meshes, volumes, etc. I’m sitting here stepping through each individual element of this level checking for CCD, opening blueprints and checking every element for CCD, checking construction scripts for dynamically created elements with CCD, etc., looking for this issue. The log is providing no help at all here, and I can’t think of why it should be that way. I would implore epic to look into that, because package-killing errors are among the hardest to track down (given that every time you want to check if the change you made FIXED it, you’re looking at another 15-20 minutes of attempted cooking) with guess-and-check.
97486-cook-2016.07.07-10.28.45.txt (3.66 MB)