fix: duplicate logs

[kyscott18]

Description

The factory child address logic was not deduplicating its results, leading to syncLogsDynamic being called with many duplicate values. This function has some address batching logic, which would cause multiple replica eth_getLogs requests to be made. The resulting logs failed when being inserted into the database.

Example

• factory contract with many duplicated child addresses

[tk-o]

Awesome update, @kyscott18 🚀

[typedarray]

Nice find. To clarify - I see two changes here:

1. When a user passes an array of hard-coded addresses in their config, de-duplicate the array while building the source/filter.
2. When fetching child addresses from the sync store, use distinct to de-duplicate the result array (because its possible for a factory to emit multiple logs specifying the same child address).

The user bug reports were caused by the factory case during the historical sync - we would fetch an array of child addresses and include it in an eth_getLogs request. If that array contains duplicates, a compliant RPC should not return duplicate logs. But, we actually separate the child addresses into batches (50?) for eth_getLogs. So, we had a situation where the same address would be included in multiple batches, leading to duplicate logs even for compliant RPCs. Is that correct?

[kyscott18]

Correct