Open charlesreiss opened 6 years ago
Did we want to implement reference counting as part of the solution to this issue? If that's the approach, does anyone want to start implementing it?
I'm not sure what you mean by reference counting.
With respect to deallocating arrays we already do reference counting...
A fix would be for the array compare functions to track the list of addresses visited so far on each "side" of the array compare and do something special (probably produce an error) when they detected a cycle.
This isn't quite just counting the number of times each array is reached during the compare because I think we want, e.g.
{ "repeated" }list array_pin var! x
{ x @ x @ }list { x @ x @ }list array_compare
to work.
You're right; my fault. I misinterpreted something you said on FM.
The MUF program
causes the server to overflow its stack.