Prerequistes

Install Node v18 and yarn (or pnpm)
Install Foundry
Create a personal access token (classic) in github with the following permissions: codespace, project, repo, workflow, write:packages

Create global .yarnrc.yml file: touch ~/.yarnrc.yml and paste the following:

npmRegistries:
  https://npm.pkg.github.com/:
    npmAuthToken: <Your GitHub Personal Access Token>

Run yarn to install dependencies
Run forge install to install other dependencies

Testing

Run: forge test. E.g.

forge test -vvv --no-match-test "SlowFuzz" will run all of the tests except some exceptionally slow fuzzing tests.
forge test -vvv" will run all of the tests

Gas costs

We have saved some gas cost snapshots, with the latest typically being saved at .gas-snapshot. To see all snapshots run ls -a .gas*.

To generate and updated gas snapshot, run forge snapshot --no-match-test "SlowFuzz".

To diff current gas costs with an earlier snapshot, pass the earlier snapshot as the --diff argument, e.g.: forge snapshot --diff .gas_snapshot.preOptimisations --no-match-test "SlowFuzz".

Deployment

TODO.

Open issues

Testing is very much a work in progress. Lots still to test.
Need to add the concept of spread to all trades, so that LPs take the spread. (Safest to do this alongside relevant tests.)
Deployment logic is non-existent
We need to integrate with the v2-core repo, and in doing so we will presumably discover some API incompatibilities.
Authorisation of many non-view functions is absent. Need a clear design and implementation here. msg.sender will probably not be the end user, but one of our contracts.
(how/when) are old positions removed from the positions list, to speed up future operations?
Measure code coverage, gas usage etc.
Linting, formatting, git hooks and CI config, etc.

Testing plan

Continue with docs and tests of low-level functions in isolation, and work up to higher functions that call them.

Docs will act as a sanity check that I'm testing the right thing.

flipTicks?
trackFixedTokens
trackValuesBetweenTicks
calculateUpdatedGlobalTrackerValues
getAccountUnfilledBases
getAccountFilledBases
trackValuesBetweenTicksOutside
growthBetweenTicks

Fix issues found.

Once we have tests, we can refactor some of the above to be more efficient. E.g. remove substantial re-fetching of the same data within a swap's tick iteration loop.

Then move on to higher level functions: