I don't know what the fvck you people are installing, but my n900 alarm has never failed, and my son hasn't been late for school as a result. And if for some reason you turn off your n900, a set alarm will turn it back on for fvck's sake!

And as far as Apple goes, I think this bug only surfaced for non-repeating alarms. So people that actually know how to use a work alarm were unaffected. I think.

Don't own one, doesn't bother me. I am curious however to know the why behind this. Hopefully it will be something that will benefit the rest of us, unless this is strictly a coding error with iOS 4.

Apparently, it is easier for Apple's engineers to produce smooth kinetic scrolling than a fully working alarm application.
Strange.

Both a simple clock and kinetic scrolling are fairly straightforward to program. Mobile devices don't have simple clocks though. They need to track time zones, daylight saving time, leap years, exceptions to all of those, and users that move between locations observing different times and rules. Mix in calendars that need to track alarms and appointments in the midst of all that, and the situation can get hairy.

As the iPhone has made news with similar clock bugs on several previous occasions, it's perhaps a bit surprising that Jobs hasn't made a complete review and correction of the clock code a high priority in his own, special way, but it doesn't surprise me that there were bugs lurking in the clock code in the first place.

WONTFIX is not decent behavior. Particularly in the case of a supposedly open-source operating system and open-source development, it is aloof carelessness at best.

So back to the topic at hand: has anybody seen any info on why this happened. The DST-related alarm glitches are pretty easy to understand, but as a programmer I wonder what perfect storm of complications causes alarms to stop working for a three day period at the beginning of 2010. There is nothing special about the unix time value (though maybe in 2038 there will be) and looking at the struct used in c's time.h, there is nothing special about this time period, either. Any clues out there?

The fact that we don't know, and probably will never know, is the reason I like open source, and why iPhone (and WinMo) will always suck compared to my maemo pocket cinderblock (and, I guess, Android as well).

Hah! Saw this posted in a Droid Fans posting on Facebook:

"IPhone Alarm Glitch Update : Until further notice Apple recommends going to sleep with someone who owns a Droid."

I kinda figured it was obvious I was kidding on that one. *Shrug* Yeah, I agree, it's pretty shitty behavior. But of the two shitty ones, I prefer the "nah, screw it, you guys deal with it how you can" to the "yeah we fixed it.... Oh! Hmmm, it seems to have happened again. There you go, we fixed it." (Of course, it depends on how much it's an honest mistake and how much of it is really negligence at reappraising their code after the first bug or two.)