playlist format UTF8 vs ASCII

[nieroster]

Winamp uses another file extension for UTF8-playlists (.m3u8)

Why doesn't Rockbox use the same file extensions to distinguish between ASCII (.m3u) and UTF8 (.m3u8) formatted playlists?

It would be nice to have a compatibility mode to read old playlists.

[bluebrother]

because Rockbox switched to utf-8 and playlists use Rockbox' encoding, which is utf-8. Recently the extension m3u8 got added for utf-8 encoded playlists. There was some discussion about changing this but it hasn't been addressed yet.

Converting a playlist isn't really hard.

[nieroster]

The problem is if I leave the extension to m3u Winamp doesn't read the playlist correctly anmore.

But if I have to change the extension anyway why not let Rockbox read all the "old" playlists in ASCII?

[bluebrother]

because nobody has implemented it yet?
Talking of pure ASCII playlists: the lower part of ASCII (i.e. up to 0x7f) is identical for utf-8 and ASCII. So as long as you don't use any extended characters (like umlauts) in your filenames there is no difference, and winamp should read those files without any problem. Possibly Winamp breaks because Rockbox uses a / as path delimiter, but that's a different story. In that case, try the iriverify plugin.

[heze]

because nobody has implemented it yet?
...

If so, why did it work until months ago?
That means, it already was implemented, but was removed. Why?

[bluebrother]

That means, it already was implemented, but was removed. Why?

It wasn't removed. It was never there! Period.

As I already said, if there are no characters outside of ASCII in your filenames there is no difference. So for most users there shouldn't be any difference. And, please read my first post in this thread. Why are you claiming thing that aren't true?  

[LinusN]

Well, Rockbox was able to handle umlauts in playlists before the unicode change. So he is telling the truth.

[bluebrother]

I don't think so. The codepage was changed, so the playlists encoding changed too. He told something was removed, but there wasn't anything removed -- the encoding was changed. Which is different IMO.
Also, using umlauts in playlists still works (I have such playlists myself!) -- they only need to have a different encoding (namely utf-8).

[heze]

Well, Rockbox was able to handle umlauts in playlists before the unicode change. So he is telling the truth.

Right! This statement confirms my logic:

Rockbox of September 2005: m3u-playlists with umlauts work fine!
Rockbox of October 2006: they have to get converted now.

Is it so difficult, to support both kinds of encoding?

[markun]

I'm working on it

[heze]

I'm looking forward, markun! Thank you very much.

When do you think is this feature available (again)?

[bluebrother]

Right! This statement confirms my logic:

Rockbox of September 2005: m3u-playlists with umlauts work fine!
Rockbox of October 2006: they have to get converted now.

Strange logic.
I'm running my linux box utf-8 only since some years now, which also results in e.g. xmms (the linux counterpart to winamp) to produce utf-8 playlists. Which means:
- Rockbox of Sep 2005: playlist with umlauts work fine. Files use latin encoding
- Rockbox of Oct 2006: playlists with umlauts work fine. Files use utf-8 encoding
So telling "they have to be converted" isn't completely correct -- my playlists (that were utf-8 before) needed to get converted until utf-8 was implemented. "They have to be converted" doesn't apply in all cases, so this isn't completely true. It all depends on the local encoding you're using (try opening a utf-8 file with notepad, which can't handle utf-8). Text files don't have any information about their encoding attached so a text file could be any encoding.
I'd rather say "Playlists need to be in the correct encoding to use characters outside of ASCII".

I disagree with LinusN as Rockbox is still able handling umlauts in playlists. The problem is simply that there is an assumption of the encoding. Which doesn't match in every case.

[heze]

I'm working on it

Hi markun,

how are the things now? Did you solve it?

Sorry, I did not check out a newer build...

Thanks.

[markun]

Sorry, I didn't finish the code yet. First I wrote it in the wrong place (wrong functions) and then I got too busy with the rockbox port to the Toshiba Gigabeat (still am)

I still have the code and should give it another go soon.

[Lear]

I've just written something that looks promising in the simulator (that's why I asked you about iso_decode on IRC ), and I should be ready to commit it in a few days, once I've had the chance to test it on target.