Post by Max KellermannNo, it just remembers that the playlist file exists (ignoring its
contents), and will tell that to the client via "lsinfo". The client
can then choose to issue a "load" command for this playlist file. Or
it can use "listplaylist[info]" to view the contents of the playlist
file.
Thanks for the info. I now understand that the mpd will treat it as any other
file in the music directory. It simply keeps it in it's database because the
file exists. The only difference being that it will not have any tags (scope
identifiers) defined, except 'filename', right?
Client: now I'm trying to understand what a client should do.
So imagine the client retrieves the db at startup with listallinfo then it will
get that playlist as
'file: x/y/rasio/radio1_Stream.m3u'
without any further tags, because there are no tags defined on that file.
The client then should add the file to the db-cache just like the other files,
but without tags.
Then the user can only find the file from it's filename (using the db-cache or
'search'), which is fine I guess. He will issue a command as if that 'file:' is
a song: 'add x/y/radio/radio1_Stream.m3u' (or addid) and the command will fail.
Right?
So the client should see for himself that the song 'file:
x/y/radio/radio1_Stream.m3u' is actually a valid playlist format/extension and
issue 'load x/y/radio/radio1_Stream.m3u' So the 'playlist intelligence' must be
in the client, right? Isn't this strange? Shouldn't the mpd also respond to the
'add' command for transparancy, to avoid client side intelligence?
recursive:
Assuming the client does the above 'load' correct, then what will happen at
'load' if the playlist itself contains another playlist?
stream:
One final question:
The lsinfo or listallinfo will return 'file', 'directory', 'playlist' (for
internally stored playlists). But somewhere I seem to have seen 'stream:'
output. Does that exsist?
thanks,
Christof