Hi!

I use amarok since a long time. I used to compiled it myself without any problem. Lately, i decided to use amarok-1.4 svn and loved it. But there\'s a problem with this version that doesn\'t appear with 1.3.7 version. Since i\'m french (from quebec), i have many frenchs songs with accentued caracters (à, é, é, ...). Randomly, amarok replaces some of these letters with stranges symbols in the database (and the playlist). Are you aware of this problem?
The problem appears also with mysql and postgresql, so i guess it\'s not sqlite related. Maybe there\'s a problem when reading id3 tags.
Juste for the info, about all my music are in ogg (i did encode it myself with grip). Tags are ok when read with other apps (like juk or konqueror).

I\'d like to help to find the solution, but since i\'m no developper, i don\'t know where to begin... Thanks!

I\'ve recompiled today\'s svn version, but using make instead of unsermake and the problem is gone!
The 1.3.7 version i was using were rpms rebuilded from cooker (i am using a very modified Mandriva 2006 version)
Thanks anyway for the help!

Doh!

The problem\'s still there, even with make. The problem appears when i ask amarok to rebuild the database. I tried some tricks, just to be sure:
--> i\'ve recompiled svn version using --without-included-sqlite option (and using threadsafe sqlite 3.2.7) and it changed nothing
--> I\'ve tried recompiled svn version again using --enable-final, but there were an error compiling

So, the problem appears only in svn version, using mysql or sqlite (the problem seems to be worst when using mysql). You have to note that french song\'s tags aren\'t not changed by amarok; the\'re only misread when creating database.
Maybe this is unicode (utf8) or ISO-8859-1 related problem. I didn\'t not activate unicode on my mandriva setup... I\'ll start looking in the sources for a hint!

Could it be that you are using MP3 files with ID3-V1 tags? Note that ID3-V1 does not support different encodings, it\'s all just ascii text.

Back in the Airborne series with had a recoding feature called \"Do not encode the following as blah.. \". The recoding features were removed in 1.4. If you have trouble with the encoding of MP3 files, you should convert them to ID3-V2, which uses utf-8.

Back in the Airborne series with had a recoding feature called "Do not encode the following as blah.. "

And why was this function removed in 1.4 (SVN)?
I have as much as 2000 songs which tags are encoded in cp1251 and now they don\'t appear in Amarok as they should be (and as they appeared up to version 1.3.7).
How I\'m I supposed to recode in utf-8 all those 2000 songs\' tags?
I think this function is needed by all that don\'t use English and utf-8 encoding,....
Please bring it back in the release 1.4 version...

Post edited by: PhobosK, at: 2005/12/19 13:27

All my oggs (since i don\'t have any mp3s) are encoded with grip and used ID3-V2. The problem seems to be related with something else that is random. I say that because sometimes there are 50 ogg\'s that are misread by amarok and other times, it\'s 120 others ogg\'s. Sometimes, all my database is ok.
I\'m getting a little bit lost with this problem!

Thanks for the help!