I spent yesterday testing every possible PVE mesh inclusion and subsequent game packaging scenario, as well as “all” known fixes (to include a baker’s dozen of my own), as sometimes (usually close to creation or upon creation and first placement of a new PVE export skeletal mesh), the package is good and playable. But, the occurrence of a “good” packaging is more likely random with lousy odds.
OCCURRENCE TEST SCENARIO
Once I realized the issue was discretely PVE mesh-related (exporting to static mesh and opening the static mesh to CPU are NOT fixes), I tested the likelihood of occurrence.
As I was able to launch a fresh-project game with three instances of a skeletal mesh tree in-scene, I then placed each tree in its own sublevel. I then packaged the game to three empty, numbered folders and checked the game logs of those that crashed.
FINDINGS
All crashed! But they did so in different sublevels.
I then repeated the test three times, and never had another success. But, the game crashed, covering every combination of sublevel, in which the asset had crashed.
This means that the problem is random with a low likelihood of success after first placement of a newly exported PVE mesh. And it’s always the same proceeding “198” line in the log:
LogLoad: LoadMap: /Game/FirstPerson/Lvl_FirstPerson?Name=Player
Assertion failed: At >= 0 && At < (uint32)IndexStorage.Num() [File:D:\build++UE5\Sync\Engine\Source\Runtime\Engine\Public\RawIndexBuffer.h] [Line: 198]
-- followed by:
Assertion failed: At >= 0 && At < (uint32)IndexStorage.Num() [File:D:\build++UE5\Sync\Engine\Source\Runtime\Engine\Public\RawIndexBuffer.h] [Line: 198]
IMPORTANT NOTE AND CAVEATS
The specific sublevel that did not load preceded the lines above in every case, failing there in the loading of that sublevels mesh.
I loaded levels in the same sequence, every time. Thus, if the first or second sublevel crashed the game, the viability of cascading meshes is unknown.
Also, the issue may be, specifically, related to the mesh’s nanite foliage.
------------