New Aiken test for process_orders with 30 shuffled donation orders

[francolq]

No description provided.

[francolq]

I made a mistake in the commit message, they are donation orders

[francolq]

Curiously, the test is giving the following result:

│ PASS [mem: 17142175, cpu: 6699341695] process_30_shuffled_orders_test

Way above the memory limit of 14 million.

[Quantumplation]

That's not surprising; 37 is our maximum batch size, but that relies on the orders being sorted in-order. In practice, they'll be shuffled, which will lower the number of orders we can fit in a batch (though it will still be much higher than v1).

Additionally, we allow the scoopers to batch orders placed before the market open time in an arbitrary orders; this means we can randomly select orders, and index them in the on-chain sorted order, allowing us to burn off the initial market demand in the fastest way possible.

We'll need to put some care into biasing our selection slightly to higher transaction IDs to offset the advantage you'd get by having a lower transaction ID, but this is something that can be solved for outside of the contracts.

[Quantumplation]

@francolq by the way, thanks for the pull request :D