I've resolved my actual problem, but it took me a while to debug, so I
thought I would bring it here in case others were affected.

I was having a bear of a time getting rid of duplicate instances of the
same song.  I rescanned the database, I wiped cache, I deleted the
database, but the duplicates kept coming back.  I finally tracked the
problem to old playlists that were pointing to an alternate path to the
same files.  I didn't realize that the scan looked for the files
referenced in a playlist even if they aren't in the music path defined
in slimserver.

Here is my setup:

My music library is divided into 3 subfolders, "flac", "mp3" and
"small_mp3".  "mp3" is filled with HQ mp3s and is where most of my
music is stored.  Recently, I decided to scan new and favorite CDs as
flac.  I made the "small_mp3" folder as a low quality mirror of the
flac directory for convenience with my portable player.  

I then put a symlink under both flac/ and small_mp3/ to point to the
/mp3 directory.  This way I can point to either /flac or /small_mp3 and
see 100% of my music in the best available format for the device.  So
any file in /mp3 could be reached through 3 methods.  /mp3/file.mp3 or
/flac/mp3/file.mp3 or /small_mp3/mp3/file.mp3.

Originally I had only the /mp3 directory, so my old playlists pointed
at /mp3.  When I changed my library structure, I wiped the cache and
thought all was good.  Eventually, I started seeing a duplicate track
here and there.  It was a minor annoyance that just wouldn't go away. 
When I updated the old playlists, I could then wipe-cache and the dups
disappeared.

Possible way SlimServer could have helped me:

Highlight tracks that are not in the official library path (but do
exist).  
A)  Possibly with a comment in the Track Information screen that points
out how slimserver was aware of the file.  I kept beating my head
against the wall because I could see extra instances had a path through
/mp3 instead of /flac/mp3 and I couldn't figure out why slimserver was
jumping to a different branch of my file tree. 
B)  Recognize that 2 different paths to the same file are somehow
unique.  I don't know if this is possible through the OS, but filesize,
timestamp and tag data should be pretty unique when taken together.  

BTW:  6.1 and now the 8/20/05 nightly of 6.2 have been rock solid on my
Debian/Sarge, wireless and occasionally wired, SB2 for months now.  Skip
free, crash free, integrated with MMM, running without intervention. 
Thanks!!! 

Apologies for the length.  

Chris


-- 
wactuary
_______________________________________________
beta mailing list
[email protected]
http://lists.slimdevices.com/lists/listinfo/beta

Reply via email to