[batcher] Cleanup batcher channel inclusion block logic

[mdehoog]

Description

This PR removes the confirmedTxUpdated bool, instead just calling updateInclusionBlocks eagerly rather than lazily from within the isTimedOut() and isFullySubmitted() functions. This has no performance implications as those methods are called at this point anyway. The isTimedOut() and isFullySubmitted() functions are getters so it's nice to clean up any struct mutation from these functions.

Tests

No logic changes, existing tests should provide coverage. Updated the TestChannelTimeout to better test the timeout by calling the TxConfirmed method.

Additional context

This is a small cleanup from #12337 that I wanted to cherry-pick in, even though that PR was closed.

[geoknee]

I'm on board with the direction of this, but I'm wondering why we need to store the inclusion blocks in the channel struct at all.

What if isTimedOut just computes these values on the fly? If they are needed at any other point (I don't think they are), we could have another getter method for them.

[mdehoog]

What if isTimedOut just computes these values on the fly? If they are needed at any other point (I don't think they are), we could have another getter method for them.

Happy to add dynamic getters for them, but note they are used in the logs in TxConfirmed and so we'd probably want to cache them on the struct (so we don't loop through the confirmedTransactions each time), which would end up looking pretty similar to what we already have