Current file processing implementation will process all files at once

[dlongley]

We probably want a processing queue instead to limit resource usage / potential main thread blockages.