Hi,

I noticed that when amarok scanned my collection, it dropped all the files/directories containing latin-1 characters.

My files are well organised. I have a directory per artist, in it I have a directory per album and in that I have the various songs that make up the album. Here is what I found

          None of the albums under the "Mylène Farmer" directory show up

          Under the directory "Laurent Voulzy" I have 2 albums: "Saisons" and
          "La Septième Vague". Only "Saison" shows up in the collection

          In the album "saison", the song "Surfin Jack.ogg" shows up, but
          "Désir Désir.ogg" does not.

So it does look as if there is a problem with the way file names with Latin-1 characters are scanned.

I use Gentoo, SVN revision 969340

Cheers,
            François

hi françois

i have files in my collection such as:

    /home/media/music/Sigur Rós/Ágætis Byrjun/Sigur Rós - Ágætis Byrjun - 07 - Viõrar Vel Til Loftárasa.mp3

which are all listed and play as expected (this is using amarok-2.1 rev. 969708). so it looks like there may be something wrong with the files in your collection.

you say that none of the albums under your "Mylène Farmer" directory show up - but are there any music files in there which have tags *without* latin-1 characters in them?

if there aren't, try putting one in there which has just plain ascii characters in the tag and do an update. does it show up?

if it does, it may be that there is a problem with how your tags are written: what program did you use to write the tags?

i use kid3 to write mine and have never had any issues with it. try using kid3 to remove all the tags from a copy of one of the problem files and then re-tag it using id3v2 only. does it show up in your collection, now?

if not, come back with more info about your general set up.

HTH

Also see this bug report:

https://bugs.kde.org/show_bug.cgi?id=189527

Thanks for the answers.

Baswal:
I took those files/directories as example. All my files containing those characters in the filename are missing, this includes a lot of French and Brasilian songs.

All my files are essentially tagged correctly (UTF-8) I used "JuK" and "rip" to tag them and they all show(ed) up when using Amarok 1.4

Note also that 97% of my files are Ogg.

Note also that I have a number of files with Japanese, Korean and Thai tags. Those are under directories/names with names containing only ASCII characters and so they show up correctly.

I did a little experiment. I took one of the file that was missing and copied it to a file with only ascii char in its name. I "Update Collection"'d and the file showed up with the correct (accented) title. From this I conclude that the problem is not with the tags (since they show up correctly) but with the filename.

Mark: Yes it  does seem to be the same bug... Since it was working with Amarok 1.4 (and I was using a MySQL backend)... The fix can probably be found in the older version. Or you could, if you only use them to reference the file, base64 encode/decode the filename in and out of the database. 

Regards,
        François

hi françois

glad you've made some progress in diagnosing your problem

my suggestions were trying to establish whether it was the tags or the filenames which were the cause and it looks like your experiments have proved it's the latter. however, this still leaves my first point unexplained: i have some filenames with non-ascii characters in my music collection and so do you - but mine are all handled correctly by amarok, and yours are not.

just as a matter of interest, what encoding is used for your filenames? (mine is utf8).

also, here's what i get if i run "mysql --print-defaults":

    mysql would have been started with the following arguments:
    --user=root --port=3306 --socket=/var/lib/mysql/mysql.sock --no-auto-rehash --default-character-set=utf8

presumably, your mysql default-character-set is different from your filename encoding.

Hi Bazwal

bazwal wrote:just as a matter of interest, what encoding is used for your filenames? (mine is utf8).

I had guessed you used UTF-8 for you files encodings... and that is why you have no problem... As for me... as indicated in the subject , I use latin-1.

The problem as explicited in the bug linked in Mark's post, is that the tags are in UTF-8 and so is the database default colation, so if your filenames use a different encoding, you run into trouble when trying to store them in the DB.

Anyway, that how I understand it. 

I do not know the intricacies of Amarok, but if the stored filenames are only used to reference the files (i.e. never used in a db query or only used in full, no LIKE, REGEX and the like), you probably could base64 encode/decode those filenames in and out of the database and never have to worry about their encoding again...

Just my 2 cents ....

Cheers,
        François

ah, sorry - to start with, when you said "latin-1 characters", i thought you were just referring to the accented characters in your filenames. for some reason, it didn't occur to me that your system would be set up to use latin-1 - i suppose i just assumed everybody used utf8 these days.

anyway, hope your problem gets fixed soon!

regards
baz