Weird jit.turtle behavior
I am trying to figure out how the pen thickness commands work inside of [jit.turtle] branches, and I've found some behavior that I can't explain (shown in the patch below). I would expect the branch stack to either ignore pen state (saving only turtle position and heading when a branch is created) or keep track of the pen thickness and restore it when the branch ends. But I don't think the behavior of this patch is consistent with either implementation. Am I missing something obvious?
[jit.turtle] also seems to have problems with negative coordinates, in the sense that operations which should be reversible are not. The patch below demonstrates that something which works in the center of the LCD stops working when one or both coordinates becomes negative.