Seems that database initialization occurs in one step now, going straight from building to committing rather than requiring a reboot in between the two phases.
Is this an intentional change? Doesn't seem to be a problem (though I can't really tell if it's missing anything in the database), just wondering if it was intentional or an accidental effect of another change?
Incidentally, I gather that devs don't bother using the database, but personally I find it very useful, as navigating by file structure is tiresome for podcasts when they often have 'artist' fields that change from one episode to the next, that make it non-obvious what folders any given podcast has ended up in. Plus I like the database custom config feature. It also works better for compilation albums where album-artist is 'various'.