[Closed] Is there a recursion limit in maxscript?

Oh, I see. Stack overflow, I would expect. But at first sight, the stack on your test recursive function seems way too small, and you would expect a descriptive error message for such a predictable error condition. I think it would be worth AD having a look at it.

Yes, that would be nice.

They have been having a look at it, which is where I’ve been getting my information. Hopefully they can at least stop the crash from happening. But even if the crash gets fixed, looks like recursion will still not work on large data sets.

• Neil

If I execute Neil’s sample script I get to 1327 recursions before Max crashes. I noticed that Neil is passing in theValue each time by value, which means each iteration of testFunc is allocating four bytes off of the stack to store theValue. I rewrote the sample using reference passing:

fn testFunc &thevalue =
  (
  	if thevalue != 5000 then
  	(
  		thevalue += 1
  		sleep 0.01
  		print thevalue
  		testFunc &thevalue
  	)
  	-- no need to return since we're passing by reference.
  )
  	
  thevalue = 0
  testFunc &thevalue

That way thevalue only ever takes up 4 bytes. Not 4 bytes * # of recursions. Which, over 1327 iterations equals 5,308 bytes.

When I ran that I discovered two things, neither of which I expected:

1. MaxScript bailed at 1019 recursions, far fewer than I expected.
2. Max didn’t crash. MaxScript generated a ** system exception ** for testFunc(). According to the listener there were 11 stack frames that bombed at the time. Each stack showed thevalue as being 1019, so the reference passing was working correctly.

I can’t explain the results. Max no longer explodes which is nice, it is presumably using less memory than Neil’s original example, but will only execute 76% of the number of recursions as the pass by value version of the sample. Based on the number of recursions of the pass by reference version I’d be inclined to say that MXS has a recursion limit of 1024. But, the pass by value version goes to 1327 which pokes a lot of holes in that theory. Even with the pass by value version the multiple allocations for thevalue only eat up 5308 bytes of the stack by the 1327th recursion. That’s about 1/4 of 1% of the stack. I can’t imagine the crash is due to the stack being 100% allocated. Something else is causing the system exception and crash behaviors.

shrug