test: Executor.Close() persists snapshot to disk

[ARR4N]

Closes #29, an open question after merging the saexec.Executor.

[StephenButtolph]

I'm not sure that this closes #29. Or perhaps I had really wanted a more general issue. The comment I made on the executor PR was intended to be a bit more holistic than just the snapshot logic.

Specifically, (after talking with @alarso16). I think we need to be calling triedb.Commit to flush the trie nodes to disk during shutdown. Additionally, I think we will want to periodically be committing the triedb during execution to limit the amount of reexecution that is required upon a restart (in the case of an ungraceful shutdown).

[ARR4N]

I'm not sure that this closes #29. Or perhaps I had really wanted a more general issue. The comment I made on the executor PR was intended to be a bit more holistic than just the snapshot logic.

Ahh your DM said snapshot.

Specifically, (after talking with @alarso16). I think we need to be calling triedb.Commit to flush the trie nodes to disk during shutdown. Additionally, I think we will want to periodically be committing the triedb during execution to limit the amount of reexecution that is required upon a restart (in the case of an ungraceful shutdown).

Yup, we are absolutely not persisting everything to disk yet. If you look in the PoC that's done in the maybeCommitTrieDB() function. It may turn out that we move that into the Executor, but it definitely wasn't my intention to include it in the last PR (hence your confusion).