Le mardi 26 avril 2005 Ã 19:07 +1000, James Livingston a Ãcrit : > On Tue, 2005-04-26 at 18:32 +1000, Nicholas Gill (mythagel) wrote: > > I've just had an idea, one that would go a fair way to solving these > design problems: multiple RhythmDBs. First off let me start by saying > that I'm not sure whether RB's design has had there being only one DB as > an assumption or not, so this might be A Lot of Work(tm). > > You would have the library/standard playlists working as normal, dealing > with music that is on your computer (using the normal tree/gda DB). > > When you plug an mp3 player in it would show the expected new source, > but instead of possibly adding the songs to the normal library (or not) > it would use its own RhythmDB (a mp3player backend). If it was a player > that supported playlists (e.g. an iPod) it could show the playlists > backed by the player's RhythmDB. > > If you dragged songs between sources that use the same RhythmDB (e.g. > the iPod playlists above) it would work like normal. If you dragged > songs to a source that was using a different RhythmDB it would copy the > song. For most things this would make sense (i.e. copy to the mp3 > player), if you dragged one to the "on my computer" library it would > have to ask you where to save it (or something). >
Depends on what you call a RhythmDB. What the iPod source does is to fill rb database engine (what is implemented in rhythmdb/rhythmdb.c) with data it reads from an iTunesDB file (the binary iPod database) instead of getting the data from a rhythmdb.xml file. Dunno if that is exactly what you want, or you had something else in mind. Christophe _______________________________________________ rhythmbox-devel mailing list rhythmbox-devel@gnome.org http://mail.gnome.org/mailman/listinfo/rhythmbox-devel