drop separate account retrieval transaction retrieval has it integrated
stop closing acuired db handles and leave that for the application termination these aren't reference-counted and can be destroyed by one thread while another still uses them
increment background task count while retrieving accounts too
another major rework, transaction list is fully asynchronous
handle cancellation
remove context parameter from MLDB.getWritableDB
major refactor, make account summary and transaction list fragments, part of the "Main" activity
reformat/rewrap
move DB stuff into a static class downside is that we need to pass a context around
separate packages for the different aspects of the application