Non-ASCII (or non-ISO-8859-1?) characters are not properly read from ID3 tags

[Chris Wagner]

Such characters are properly read from Ogg-Vorbis files, as we are using our custom, Haskell-based Vorbis module for reading those; and this module is conscious of character encoding and assumes UTF-8. However, when reading MP3s, we're using TagLib to do the reading.

It may be possible to get the proper strings by decoding the values handed back by the C-wrapped TagLib library. However, we may prefer to just adopt the Haskell-based "idiii" tag-reading library, as that may also solve our problem of only getting basic tag values (artist, album, title, etc. -- nothing advanced, like MusicBrainz IDs) from MP3s, too.

[Chris Wagner (chris-wagner)]

Addressed in revision 863.