- don't use now() in WHERE clauses: hurts perf at least on PostgreSQL

- bug with multiple connect entries matching a reconnection (only in multiple
  SQLgrey instances configs and in a rare race condition) to fix
  -> accept multiple connect entries in "in_connect(), recently_in_connect()"
  -> remove the delete in "put_in_connect"

- external cleanup process/experiment with an adaptative cleanup algorithm

- switch from log level to log categories

- add an utility for statistics (reading the logs)

- create a stress-testing suite (for performance/reliability tests)

- use time units everywhere to make configuration easier

- check if permission related SQL errors are catchable

- try to exit with a proper error code when database is unavailable or the user
  doesn't have enough rights to access the pid-file

- consider optin/optout (is it easy to do with Postfix? Then provide an HOWTO
  else code it in SQLgrey)

- add sender e-mail based whitelist with huge warnings

- experiment with SPF support

- support migrating from another greylister by calling it to learn its auto-whitelists
