#5 Dictionary strings not encoded in UTF-8

[Jordi Monfar]

Have found that albums/tracks that contains non pure
ASCII characters fail to play.

According to Archos specifications for the library, strings
have to be encoded with UTF-8. The current jbmgen
does not have any provision to change the encoding
from that is found in the Archos files, usually if not ever
Latin1.

This simple modification encodes strings in UTF-8 when
added to the dictionary:

bash-3.00$ diff Dictionary.rb*
40c40
< data.putstring(self.unpack("C*").pack("U*"))
---
> data.putstring(self)

After introducing this change, non-ASCII strings are
found on the library encoded like MusicMatch does, and
the device interprets them correctly, playing the albums
and tracks without problems.

Hope this helps.

[Bachelier Vincent]

Logged In: YES
user_id=1094185

I have test with that one:

data.putstring(self)

from original pack, but all accent are bad encode, but I can
read from arc library selection my track

and with your patch:

data.putstring(self.unpack("C*").pack("U*"))

all accent are successfull shown in Arc Library, but I
couldn't read any more my file. My file have accent too

I don't program with ruby, and I just want to know how can I
do to just encode mp3info, and not filename in utf-8

To know:
My system are totally under UTF-8
I use iocharset=utf8 for mounting vfat drive,
all done when I don't use Arc Library,
but fail with it like in the top of this message

Do you have an idea ?