In limited scope tests, just trying to recreate this problem, I can’t. So it’s something about my specific implementation. But in my game, I have a gun sprite that is placed at same position as the player with extracted properties. Once per frame it rotates the gun towards the current mouse position. Sometimes the player and/or the gun will be flipped depending on which way they’re facing. I have a much wider sprite of a red line that is attached to the gun and has ‘follow rotation’ enabled. From 12 oclock counterclockwise to 6 oclock works fine, but the other half of the 360 does not. The red line stops rotating along the right side of the circle even though the gun it is attached to turns correctly. The red line itself has almost no behaviors defined, it just fluctuates opacity.
It appears the rotation is being set to -180 <-> 180, and this is part of the issue? Unclear. For now what I do is: in the red line behaviors, I use an Always along with an Extract and an Expression to get the rotation, add 360 to it, and then set that as the objects rotation. Its unclear which of the above actually corrects the issue but it does then work as expected.
I really wanted to make a dead simple test of this to show it in effect, but unfortunately if I create just the minimum blocks necessary and setup a test, the problem does not occur.
The ‘flippedness’ of the object may be a factor, I think in the cases the red line worked, the gun was currently flipped. Then when it did not work, the gun was back to original.