pdobacz

Why not use the blocks stored in `leveldb` for this instead? Postgres (WatcherDB) might hold a filtered/truncated set of transactions, and as such should not be relied upon for recovery....

> `apps//omg_watcher/test/omg_watcher/integration/invalid_exit_test.exs:` some got added to cabbage in #1344. > Shared state between tests. The tests above run in a way where the setup of the test is basically a...

@InoMurko > @pdobacz question: the #1344 added some test both in cabbage and in elixir-omg. are we able to delete any tests that got implemented as cabbage tests now? It...

> ahaaa, okay, I only took a quick peek at the tests. are we able to identify and mark potential tests that are implemented in both "old" way and cabbage?...

I'm supportive. Note that the relations between utxos and eth events can be very different. Are we're aiming for this to read, that `events` are just events that "somehow" relate...

> Notice that the deposit event has eth_hash but transfer event has plasma_hash. Might be something to improve but we havn't got a better idea yet. I'd say, let's not...

> Also I'm not sure if the "status" field for each event makes sense. Since those events should already be confirmed by all the thresholds we set on EthereumEventListener? yeah,...

> Idea I'm playing around is if it makes sense to use the [ecto changeset](https://hexdocs.pm/ecto/Ecto.Changeset.html) this in particular is on my mind since long, never got to it :cry:

It might've been resolved by the lazy-loading of UTXOs. Unfortunately `613-state_exit_utxos_race` is no longer out there (I shouldn't have been lazy to push a synopsis to a branch and not...

We'll asses and re > if it's a mainnet blocker. see original post. Unless anyone can think of a way this can happen - the problematic condition can't occur with...