Well, it would help if the interpreter is also fussy on, e.g. not reading multi-line comments. I'm losing the motivation to stay up at 1am bug fixing what's otherwise perfectly fine code, except that your interpreter ignores the entire file.
You ain't getting nothing until you make me a proper debug tool.
Have you checked the BulletScriptErrors.txt log? That's usually where I die.
I've been experimenting with how good the angular and time precision is:
I'm trying to make bouncy snakes. Shots that follow each other with tight precision in an S curve pattern that ricochet off walls and then home in on the player. I had given up before because I couldn't keep them in line and thus they deteriorated into a swarm of bees. (Swarm of bees or school of fish shots...). Then my hope was rekindled after the boomerangs. So I tried and failed and then decided to try a precision test.
Procedure
Setup 7 trials, each of which should take the same length of time to complete an 180 degree turn.
Loops # of loops performed
Angle = Degrees change per loops
Time = How long each loop step should take
TorchWhite Loops=1 ANGLE=-179 TIME=1.8
TorchRed Loops=170 ANGLE=-1 TIME=0.01
TorchEmblazonedYellow Loops=90 ANGLE=-2 TIME=0.02
TorchYellow Loops=60 ANGLE=-3 TIME=0.03
TorchBlue Loops=45 ANGLE=-4 TIME=0.04
TorchGreen Loops=36 ANGLE=-5 TIME=0.05
TorchPurple Loops=30 ANGLE=-6 TIME=0.06
TorchPink Loops=25.7 ANGLE=-7 TIME=0.07
Notes:
Loops=25.7 I just noticed I asked for it to loop float number of times. I'm not sure what it did, probably rounded down.
Why 179 instead of 180? 180 could go either clockwise or counter-clockwise. I've had it even happen with values in the low 170s. This seemed to work fine here.
Also, my results may be related to system load or system specs, which means they are only valid for me.
Discussion
Only the 179 turn made it. The 1 degree turn just gave up and went straight. I don't think it is a take_pants_off_head problem.
Trying again with large angles.
Changed to angles 8-14, with respective changes made to loops and time.
Discussion
They do better. After 2 S turns even the 179 degree turn shot stream isn't in single file any more.
Conclusion
Keep angles above 15 and with rotations longer than 1.8 seconds if you really need precision. Tight, high precision bouncing snakes and boomerangs probably aren't going to happen very well. Regular jumping snakes will have to suffice. Invest in bees.
Pictures:
Test 1: Red gives up
Test 2: Everyone is together at first
Test 2: After 2 S curves
Test 2: After more than that S curves