feat(stdlib): Optimize List.init to support large lists

[spotandjake]

This pr optimizes List.init to support large allocations without overflowing the call stack.

On lists smaller than 2500 we use the old approach and for anything else we use a tco approach build the list and then reverse it.

Notes

• 2500 is a rather arbitrary choice to switch over, We start having stack overflows with the grain runner around 5k elements and I figured other runtimes might have smaller stacks so it probably makes sense to switch over before that.
  • If we choose to low a number we leave performance on the table because of the List.reverse overhead
  • If we choose to high a number we could overflow the wasm call stack.
• feat(runtime): Optimize == for lists  #2247 needs to be merged before this for tests to pass.