Improving the wake-up alarm

[jswetzen]

I have an iPod 5g but there is a problem with the wake-up alarm: if you happen to switch on hold before the alarm goes off, the OF will start and there will be no alarm! I was wondering if anyone else finds this irritating, and if this is an issue on other targets. Is this something to work on?

EDIT:
There is now a first patch on the tracker: FS#9087

Testers are needed, especially for other iPods than the video.

[nls]

This comes down to the bootloader not being aware of what triggered the power on and is so on all targets where hold controls which firmware is loaded.

[jswetzen]

Thank you, nls. I managed to make the iPod 5g bootloader start rockbox if the alarm is active, but it doesn't check if the alarm actually triggered the power on. Unfortunately, this wasn't enough since all settings are cleared if hold is on during startup  but that was quite easily fixed too.
I'll make a patch and upload it to the tracker, it might work for other iPods than the iPod video.

[Llorean]

Well, it needs to be able to determine if the Alarm is what woke it up, otherwise setting an alarm prevents you from booting into the OF doesn't it?

[jswetzen]

There is still the option to boot into the OF in the old way, by holding down the MENU button, but it does make it more difficult for new users.
If I can figure out the assembler code that makes the alarm work, I will find a way to check this properly.

EDIT:
The patch works well with the iPod 5g, but I really need testers for all other iPods if this is ever to be included in svn.