Add checkpoint_active on the WAL

[blakesmith]

This new call immediately sends the active log file to the background checkpoint thread, regardless of its current size.

This is useful when you need to force written log entries to be immediately materialized in the database, perhaps during a node drain, or before a shutdown event.

Addresses: #10

Tests: Unit.

[ecton]

This looks good at first glance, thank you for the contribution! It's been just long enough since I worked on this crate last that I want to check through it again before I do a release with the new feature.

[blakesmith]

@ecton Sounds good. One thing that could be improved: This change holds the Files lock twice. Once to acquire the active file, and one to pass it to the background checkpoint thread. If you would rather that be combined, and carry the mutex guard across, I can make that change. If not, I'll keep it as is.

Let me know if you need anything else in the meantime!

[blakesmith]

@ecton Any updates here?