Hacker Read top | best | new | newcomments | leaders | about | bookmarklet login

Nitpick: there's a big difference between an app crashing (which is what this sounds like) and an app crashing your phone (rendering it unusable or forcing a reboot).


view as:

There's reports of users' entire phones crashing and rebooting because of the app.

The app actually is crashing phones. It turns out that it's quite possible for apps to crash most Android phones, either by making the phone so unresponsive that a watchdog kicks in, or by using so much RAM that the OOM killer goes to town.

I stand corrected - that is pretty bad.

If it's possible for an application to bring down the system that sounds like the system was poorly designed.

And... you've essentially discredited windows, linux, and os x.

Really? Not once have I had an application take down GNU/Linux. Sure an application takes down X once in a blue moon, but never the whole Os. And I've been using it for about 15 years now. The few full scale crashes I've experienced were due to hardware errors.

Only once have I experienced a full crash on android. The phone simply rebooted. No idea why.

I have never used OSX or windows after XP, so I can't say anything about them.


I'm not sure if this is meant to be a serious comment, but not only is it misleading, it's also naive. As you try to lead the conversation into talking about an actual kernel halt or panic as opposed to what the OP actually said.

This was the original claim that you responded to. "by making the phone so unresponsive that a watchdog kicks in, or by using so much RAM that the OOM killer goes to town."

Crashes obviously happen on linux systems all the time. That's why this exists http://linux.die.net/man/8/watchdog and why we have reports like this http://www.unix.com/aix/106921-process-crashing-aix-due-memo... and this http://www.linuxjournal.com/article/6059 etc. Are you claiming that such a situation can't arise on linux.

The same is obviously true on Windows and OSX.

Pivoting the conversation isn't the best way to make your point.

edit - Not to mention you attempt to use anecdotal evidence to support a general claim. Such has no place here.


First of all: Sorry for the late reply. I somehow didn't notice your post before now.

I was actually replying to jbrennan who said "If it's possible for an application to bring down the system that sounds like the system was poorly designed." and burrows who said "And... you've essentially discredited windows, linux, and os x.". I didn't notice what ootachi said until you pointed it out. I consider "bring down the system" to main exactly that, kernel halt or panic.

The links you provided where about processes crashing and/or processes with memory leaks, which I could in no way consider to be "bringing down the system" (unless it's a very poor system where such things actually do bring down the whole system), and it would be completely insane of me to claim that processes cannot crash on nix.

My point was simply that kernel panics or anything that "brings down the system" very rarely occur, and normally occur due to hardware errors on nix systems, and as far as I know on newer Windows systems as well (again, never used OSX so I have nothing to go on there).

Simply put: When jbrennan said "bring down the system" that, to me at least, means anything that the _only_ recovery from is a reboot of the system, no more, no less, and that is what I was responding to.


In both cases, iOS would just kill the app. Why is Android not doing the same? Via limits.conf/ulimit/setrlimit, this should be quite easy to do.

Even on iOS, it is surprisingly easy to kill the springboard from an app.

Legal | privacy