Version: 1.13 (using KDE 3.4.91 (beta1, >= 20050910), compiled sources) Compiler: Target: x86_64-linux-gnu OS: Linux (x86_64) release 2.6.13.2 After downloading the freedb (the 500Mb tarball) and unpacking it in $HOME/.cddb I see that the encoding kaudiocreator uses does not match that of the download. For CDs that have characters above 127 there are always two charactrs missing in the disc info kaudiocreator shows; leading me to think that it assumes utf8 locally. Example: I see Bj?k Which should be Björk (note the missing r in the first). I suggest to check for validity of utf8 and if there are illegal utf8 chars fall back to the encoding freedb uses (sorry not sure what that is).
Created attachment 12886 [details] Example freedb file Please use this file as an example which should load correctly when placed in the local cddb database by the user.
This file is Latin1-encoded. I wonder how freedb does when you have a non-Latin1 character (like Tatu's Нас не догонят). Obviously it can't be represented in your file. I also don't see any obvious encoding markers.
Change product
I guessed that when freedb added support for utf8, they changed the files to utf8 too since that's what it sends from the server. But it looks like the server does runtime detection of the encoding and converts it if needed. I've tried before to add some kind of automatic encoding-detection before, but I didn't get good enough result with the Qt/KDE-methods to detect it. But with the code from the cddb server I seem to get good results, but I need to do some more testing. I'll do some more testing on that code, and then I'll try to incorporate it in to libkcddb.
*** Bug 119616 has been marked as a duplicate of this bug. ***
Lets close this one now :)