Notices

I've looked around how to read that button on the N900 and got very disappointed. Nokia actually thinks that nobody would ever want to use that button for other things than to answer the phone. So they hardcoded it into the OS (into that f****** policy daemon) that the button does only generate events while in phone mode. Nokia has some really bright people.

qwerty12, before he abandoned Maemo, found a way to hack ALSA to circumvent this stupid restriction. Maybe I can make use of his code.

The other way would be to constantly poll the sysfs interface (IO watching doesn't work on sysfs). But this would drain battery. So it's not an option.

Sorry, thanks to Nokia, no short-term solution available!

And what about interaction with thge headset-control package?

Yes, I think that's the only viable option.

Skipping tracks has become more slow in the last SVN update.
It takes about 10 seconds.

Have you tried rebooting your device? Something must be going totally wrong there. I have never experienced such skip times.
Does using a different player backend like GStreamer or mplayer yield the same skip times?

You can run
to get some profiling information. Maybe this can help find the bottleneck. I somehow doubt that the bottleneck is completely on MediaBox' side.

Nice work on the rotation in video mode. And now we can type to search for a song!!!!!

ps. How about youtube video's sorted by date etc.?

That will come, too.

Ok, after rebooting I got skiptimes of 3 - 5 seconds. Are that normal values?
I have a huge playlist with about 1000 tracks.

3 - 5 seconds are not normal, either. But to help you with this I need some more information about where the bottleneck actually is. Some profiling output from MediaBox might be helpful.

Yes i also think the program generally works slower that the default mediaplayer. If that can be fixed it would be just perfect