More > JRiver Media Center 31 for Linux

fix path which was broken due to accent characters

<< < (4/4)

BryanC:
What I would have recommended, and perhaps you can still get to this point, is to sanitize the filenames using the same regex/clean in MC as on the filesystem. You have special characters in the filenames that are mismatched between the MC database and the filesystem. So the easiest thing would be to standardize case and replace all of the wonky characters on disk and in MC filename field with an underscore, at least for purposes of matching up MC and the filesystem. From there you could parse the remaining mismatched entries by hand, and then run a rename to fix casing and special characters (again, I'd recommend running through Clean() on Linux unless you want to end up with improperly nested directories or hidden files, etc).

If you really want to start with a fresh library (I wouldn't as playlist restore will be a nightmare as you're about to find out) then at least force save the MC database to tags first so it is restored on next import.

leezer3:
IIRC the way charsets are stored internal to the library might have changed in the last few versions (29/30)

Bob would have to comment, as I can't immediately find the thread I'm looking for.

bob:

--- Quote from: leezer3 on July 09, 2023, 12:49:23 pm ---IIRC the way charsets are stored internal to the library might have changed in the last few versions (29/30)

Bob would have to comment, as I can't immediately find the thread I'm looking for.

--- End quote ---
This has been changed on MacOS, MC now stores the UTF8 as precomposed as opposed to decomposed (filesystem representation) because ONLY the filesystem stores them as decomposed and either works when you give them to MacOS to search for. This also makes the database character storage the same on Mac and linux.

Any decomposed paths in the database are converted on a once time basis with current versions of MC (anything 30.0.27 or newer)

Navigation

[0] Message Index

[*] Previous page