v2.0.23

v2.0.23

Note: Previously when an API call with search type 'movie' or 'tvsearch' was made without any identifiers or category I would call indexers with search type 'search' instead because some indexers don't like that. This causes some other problems so I've reverted that behavior except for a certain list of indexers. I have hardcoded list of indexers for which the switch will be done. I'm not sure which indexers actually behave that way. So if you find an indexer where browsing the movie or TV releases (e.g. using NZB360) will return a lot of crap please let me know so I can add the indexer to the list.
Note: I've changed the java runtime that is used in the docker container maintained by me (although I actually don't want to really support that...). In my tests it nearly halved memory usage in some scenarios (199MB compared to 380MB). If this proves to be stable I'll recommend the other maintainers to use this as well.