New FMath::CountTrailingZeros function.
TSparseArrayCompact will now compact in-place and not do anything if the array is already compacted.
#codereview robert.manuszewski
#platformnotify josh.adams
[CL 2691582 by Steve Robb in Main branch]
Allow derived gameplay cue managers to not begin async loading of object libraries.
Reduce priority of async object library loads
[CL 2691509 by Graeme Thornton in Main branch]
- Added a new FOutputDeviceArchiveWrapper which doesn't try to do any special work and merely adapts the FOutputDevice API to write to a FArchive
- Switched UEngine::HandleMemReportDeferredCommand over to use FOutputDeviceArchiveWrapper
[INTEGRATE] Change 2691200 by Michael.Noland@mnoland-T2784-HDepot on 2015/09/14 20:17:53
[CL 2691208 by Michael Noland in Main branch]
I found a way to reproduce the bug easily: drag a box shape into the level in the editor. make it movable. ctrl-e to edit. remove collisions. add about 6 box collisions to it. simulate. then select the box and try to scale it.
essentially elements of SomeClass in a TArray<FKShapeElem> was storing addresses of themselves. when the TArray grew, those pointers became no longer valid. these self pointers are used as 'userdata' on physx objects. When creating a physx shape from the FKShapeElem, it was getting an out of date pointer to use as userdata. Its probably the only place that uses this particualr physx shape userdata is the scale code as it iterating over physx shapes and reaches back into what was turning out to be the wrong place to get initial shape information.
given physics is recreated on each change to collision already, we can fix this issue simply by computing this pointer/userdata on the fly, rather then the first time the shape is added to the TArray
future todo: still would want a better from-physx-back-pointing system to safeguard to prevent this these sorts of thing from happening in the first place. what if this array somehow got modified without physx scene getting re-created.
crash reports: http://crashreporter/Buggs/Show/73744https://jira.ol.epicgames.net/browse/UE-20961
#jira UE-20961
[CL 2691166 by Stan Melax in Main branch]
