Initial stress test for liquidation auction

[utdemir]

This PR adds a new e2e test called LiquidationsStressTest which creates a thousand liquidation slices.

It mainly tests the push_back function, and it seems to be well below the gas limit since we're triggering 40 liquidations as a bulk operation. We should add similar tests for other AVL functions too (split and del comes to mind).

While doing that, I had to implement a few other minor changes:

• On this test, we are running a patch version of Checker on which the protected index tracks the actual index much more closely, in order to be able to trigger liquidations quickly. To use this:

CHECKER_DIR=$(nix-build -A michelson --arg e2eTestsHack true) python e2e/main.py LiquidationsStressTest

When we end up in a state where our E2E tests run reliably, we should add this to the CI.

• Our 'ptr' type is now a 'nat' instead of an 'int'. I just felt like it suits more since it always is positive.
• Our 'liquidation_auction_place_bid' function now expects an 'auction_id'
• We have a 'current_liquidation_auction_minimum_bid' view, which also returns the auction id.

[gkaracha]

LGTM! 👌 I left a few suggestions but apart from the one relating to the potentially failing assertion the rest could be done later as well. It's great to finally have a liquidation test, thanks! 😄

[gkaracha]

Ha, that's weird indeed 🤔 Not sure if it's a pytezos bug; it reminds me of #81.

[dorranh]

Awesome! Just a couple of small comments but otherwise LGTM as well. It would be cool to add some profiling in this test so that we can see the trend in gas costs with successive calls. I'm happy to work on that next week as a separate PR though.