Output processing stages

[rakita]

There are multiple use cases where users what to do something with the output state, problem with that is there are footguns or unneeded state that are not easy to consume. The returned state reflects the internal JournaledState that in good amount of cases needs to be filtered in one way or another.

Examples:

• Filter only the changed data.
• Filter all read data.
• Nulify self-destructed account.
• Commit to the database (aka Context)

We can see a few types of stages that we can have:

• Modification applying changes to data (Applying selfdestruct to remove the footgun, and prepare for next tx)
• Transformation from internal type to something that is easier to consume ( Commit to db and return TransitionAccount)
• Filtering of the data (Filter only the changed accounts).