Remove naive solver

[MartinquaXD]

Description

The team seems to agree that the naive solver is not adding a lot of value so it's reasonable to delete its code.

Changes

Removes the naive solver code.

[fleupold]

What are the alternatives considered? I realise that currently Naive solver is somewhat useless, since it doesn't support fee=0 orders, but how hard is it to add this (it should be strictly simpler than for baseline).

Given the name of our product, I'd find it unfortunate if we are only relying on a single external solver to find obvious CoWs. Also for e2e tests I think it would be good to be able to test CoWs.

[MartinquaXD]

it should be strictly simpler than for baseline

Conceptually it's fairly easy to come up with some way to distribute the execution cost across orders. However, the code is some serious spaghetti. :/
I looked more into supporting limit orders last night and I think I found a way that could make it work but it's not pretty.

[sunce86]

Given the name of our product, I'd find it unfortunate if we are only relying on a single external solver to find obvious CoWs.

Good point. OTOH I don't think Naive solver is the one we should rely on in the future for COWs (or anything). External solvers should start supporting COWs and if they don't it's just a matter of incentive then.

[fleupold]

I don't think Naive solver is the one we should rely on in the future for COWs (or anything)

For the purpose of e2e tests and making sure the driver logic supports CoWed solutions I still see value in having a solver that finds obvious CoWs

[MartinquaXD]

Closing since there is pushback after all. Already looked into supporting limit order support but the code is pretty hard to reason about and it's not 100% clear to me where the adjustment to the fee logic has to happen ideally.