Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!

 



Forgot your password?
typodupeerror
Compare cell phone plans using Wirefly's innovative plan comparison tool ×
Bug Chrome IOS Iphone Safari Software Apple

Clicking on Links in iOS 9.3 Can Crash Your iPhone and iPad (apple.com) 100

Reader lxrocks writes: Many users are experiencing an issue with their iPhone and iPad wherein trying to open a link on Safari, Mail, Chrome or any other app causes it to freeze and crash. The issue renders any type of search with Safari as useless as none of the links returned will open. The wide-spread issue -- for which there's no known workaround just yet -- seems to be affecting users on both iOS 9.2 and iOS 9.3. Apple has acknowledged the issue and says it will release a fix "soon." There's no official word on what's causing the issue, but a popular theory with developers is that the glitch has something to do with Universal Links, a feature Apple first introduced with iOS 9. It appears some apps, such as Booking.com, are abusing this capability, causing the Universal Link database to overload.
This discussion has been archived. No new comments can be posted.

Clicking on Links in iOS 9.3 Can Crash Your iPhone and iPad

Comments Filter:
  • by avandesande ( 143899 ) on Tuesday March 29, 2016 @11:22AM (#51800073) Journal
    You are clicking the links the wrong way!
    • by Anonymous Coward

      Nah, it's just that iPhones only have a fixed memory capacity, because Apple won't let its customers have a memory system that could handle a large database, so they fundamentally broke the database by crippling its ability to handle reasonable amounts of data in order that it not eat up the limited available storage required to let the unit continue to work in general. But that's okay, because Apple's customers clearly like that kind of treatment.

      I knew a lady like that once; she just couldn't really have

      • Re: (Score:2, Insightful)

        by Anonymous Coward

        "In iOS 9, developers can take advantage of a feature called "Universal Links" to associate their apps with their websites. When their app is installed on your phone or tablet, links to those sites open up in their apps instead of in Safari as they normally would. It turns out that the app for travel site Booking.com crammed every single URL from its site into the list of associated links in its app

        OK. Stupid behavior on the part of an app developer. Fair enough.

        But there's a bigger problem here . . . this bizarre mindset of creating an "app" to do things that can be/should be done by an ordinary web browser. Why exactly do you need an "app" for Booking.com at all? Yes, I know, everyone likes to app while they app, so now it is fashionable to put apps in their apps so they can app while they app. (Yo Dawg!)

        But this is just fucking stupid, and it appears that this stupid fixation on "apps" is star

        • Re:Apple Feature! (Score:5, Insightful)

          by anegg ( 1390659 ) on Tuesday March 29, 2016 @12:52PM (#51800769)
          I myself am unamused at the proliferation of "apps" that don't do much more than a web site would do. I tend not to install them, and even uninstalled a few that I had installed because I found that I really didn't like the automatic switch from the web site to the app when I just wanted to use the web site (Amazon, for instance). Although I am not an expert in iOS app development, I suspect that an app gives the vendor the potential for much greater access to personal data on the device than just going through Safari. In some cases the app may provide for a better user experience, but keeping vendors at arm's length through the browser seems more secure to me.
          • I suspect that an app gives the vendor the potential for much greater access to personal data on the device than just going through Safari.

            My suspicions precisely. It's weird how, on the desktop, it's the web browser that spies on me. Yet on mobile, the web browser is one of the few apps I trust.

        • There are honestly quite a few websites I'd use in place of the app if it wasn't for the constant messages "reminding" me that there's an app for the same thing. Plus Android's Chrome either doesn't support the notifications API or else hardly any developers do.

          Twitter's mobile website is, at the moment, actually better than their app. But I can't get it to tell me I've got DMs or mentions without manually going into it, and every time I go into it 1/5 of the screen is turned into an annoying "BTW you sh

        • But there's a bigger problem here . . . this bizarre mindset of creating an "app" to do things that can be/should be done by an ordinary web browser.

          Yeah, why make a "Mail" app, when mail works just fine in the browser?

    • that is all.

  • by ErichTheRed ( 39327 ) on Tuesday March 29, 2016 @11:28AM (#51800115)

    It's interesting to see such a large company letting a bug like this slip by, especially in an operating system. You would think even with an Agile "ship it broken, we'll patch later" mentality, they would have armies of QA people and automated scripts banging away at every corner of the OS. Something like "clicking on any link in our bundled browser with JavaScript turned on crashes the application" seems to me like a showstopper bug.

    I'm all for getting stuff rolled out in a reasonable time frame, but core stuff like an operating system needs to be tested a lot more intensely than some social media/dating app. Not everyone is connected 24/7 with easy access to patches...the product I currently do systems engineering work for is used almost exclusively in offline environments.

    • by creimer ( 824291 )

      It's interesting to see such a large company letting a bug like this slip by, especially in an operating system.

      Otherwise known as the Microsoft model of software development: ship it now and patch it later.

      • Re: (Score:2, Insightful)

        Also known as the Apple model of hardware development. Ship it now, offer a fix in 4 months for the shoddy work, then discontinue support for your 7 month old phone and release a new one.
        • by Anonymous Coward

          Got an example of Apple discontinuing support for *any* 7 month old product? Even just one? Or are you just doing a mindless, 'they do it, too' post?

        • Also known as the Apple model of hardware development. Ship it now, offer a fix in 4 months for the shoddy work, then discontinue support for your 7 month old phone and release a new one.

          Since they still support the iPad 2 and the iPhone 4s, you are clearly talking out your ass, hater.

      • Not even Microsoft would deploy something as stupid as this.
        • by creimer ( 824291 )

          Not even Microsoft would deploy something as stupid as this.

          You never tried Windows 1.0 or 2.0? I heard those version suck donkey balls. Most people point to Windows 3.11 as being the first usable version of Windows.

    • by MobileTatsu-NJG ( 946591 ) on Tuesday March 29, 2016 @11:38AM (#51800199)

      I suspect this is an intermittent bug. Anecdotally my wife and I have been on 9.3 for at least a week or two and have had no problems. This might be one of those things that slipped by because it's really hard to reproduce.

      That said, I have not been impressed with Apple's software quality in the last couple of years. I don't know if it's because it got a lot more complicated when it went 64-bit or if it's because when Steve was here he cracked the whip a lot harder, but I've definitely witnessed a lot more silliness in the software recently. iOS 9.x was supposed to be the bug-fix version, but I ain't seeing it.

      • by Anonymous Coward on Tuesday March 29, 2016 @11:53AM (#51800293)

        The bug won't express itself unless you have an app that ignores the wild-card capabilities of the Universal Link associations, *and* has a huge number of links defined (such as the Booking.com app which did their definitions in *exactly* the wrong way, having a defined link to each hotel, rather than '.../hotel/*').

        The underlying code does need to be fixed, but the sort of thing needed to expose it is exactly the sort of thing you wouldn't expect to run across, and therefore probably wouldn't think to test against.

        There's more details here:
        http://arstechnica.com/apple/2016/03/poorly-behaved-app-causing-crashes-and-link-problems-for-some-ios-9-x-users/

        • +5 informative

        • The underlying code does need to be fixed, but the sort of thing needed to expose it is exactly the sort of thing you wouldn't expect to run across, and therefore probably wouldn't think to test against.

          Rule one of software development: users are stupid, never trust user input or data.

        • The underlying code does need to be fixed, but the sort of thing needed to expose it is exactly the sort of thing you wouldn't expect to run across, and therefore probably wouldn't think to test against.

          If your software testers aren't testing the cases people don't *normally* think to test against, then you should replace them with random non-software-tester users who will accidentally test those cases. ;)

          In other words, software testers are *supposed* to test the things you wouldn't normally think to test, at least in part of a test cycle, somewhere...

          • The underlying code does need to be fixed, but the sort of thing needed to expose it is exactly the sort of thing you wouldn't expect to run across, and therefore probably wouldn't think to test against.

            If your software testers aren't testing the cases people don't *normally* think to test against, then you should replace them with random non-software-tester users who will accidentally test those cases. ;)

            In other words, software testers are *supposed* to test the things you wouldn't normally think to test, at least in part of a test cycle, somewhere...

            And by "software testers" you obviously don't mean the developers from booking.com who didn't test the effects of their new app version?

      • by tootired ( 91527 )

        I updated to 9.3 on day one. The bug hit me yesterday (28th) at around noon.

        • Out of curiosity was this on an iPhone 5 or an iPhone 6? The reason I ask is I'm curious if this was the update they did last night, the one they rushed out because of bricking issues.

      • I suspect this is an intermittent bug. Anecdotally my wife and I have been on 9.3 for at least a week or two and have had no problems. This might be one of those things that slipped by because it's really hard to reproduce.

        That said, I have not been impressed with Apple's software quality in the last couple of years. I don't know if it's because it got a lot more complicated when it went 64-bit or if it's because when Steve was here he cracked the whip a lot harder, but I've definitely witnessed a lot more silliness in the software recently. iOS 9.x was supposed to be the bug-fix version, but I ain't seeing it.

        The issue is caused by the Booking.com app. How has slashdot picked up on this issue and not the source of the bug? It's been known for over 24 hours. Booking.com's app registers every single URL their website offers (for deep linking) instead of just booking.com This causes a crash when trying to parse the list of deep links. Apple has already acknowledged the issue and said they have to release an OS Patch to fix broken devices. Booking.com has already removed the troubled app and published a new one

    • by tlhIngan ( 30335 )

      It's interesting to see such a large company letting a bug like this slip by, especially in an operating system. You would think even with an Agile "ship it broken, we'll patch later" mentality, they would have armies of QA people and automated scripts banging away at every corner of the OS. Something like "clicking on any link in our bundled browser with JavaScript turned on crashes the application" seems to me like a showstopper bug.

      No, it's because of the way the OS works with apps (which are sandboxed).

      • huge 2.8MB table

        Hahahahahahahh! Huge and 2.8MB do not go together in the same sentence! These are phones with gigabytes of RAM. Gigabytes of super-fast flash storage, and screaming octa-core CPUs. And system crash due to a puny 2.8 megabytes table?! Is this a fucking joke?

    • It's interesting to see such a large company letting a bug like this slip by, especially in an operating system. You would think even with an Agile "ship it broken, we'll patch later" mentality, they would have armies of QA people and automated scripts banging away at every corner of the OS. Something like "clicking on any link in our bundled browser with JavaScript turned on crashes the application" seems to me like a showstopper bug.

      I'm all for getting stuff rolled out in a reasonable time frame, but core stuff like an operating system needs to be tested a lot more intensely than some social media/dating app. Not everyone is connected 24/7 with easy access to patches...the product I currently do systems engineering work for is used almost exclusively in offline environments.

      If you develop software, then you know how easy it is to fall into a "testing rut". You put together test data/round-up the URLs of some sites, and use those during your Development.

      Who would suspect that some stupid web-developer at the other end would try and shove megabytes worth of "link" data down your throat when you so much as clicked on a SINGLE LINK?

      This was an corner-case, clearly.

      • by Anonymous Coward

        It's a very well known corner case and one of the first things you're supposed to check for. When receiving any type of input, you write checks for missing input, massive input, and malformed input. If you're not always doing those things, you shouldn't call yourself a tester (really the developer should have handled all these first anyway). This is one of the most basic QA things you're taught in any decent software engineering degree. Never trust input!

        I seem to recall browsers having issues with long

        • It's a very well known corner case and one of the first things you're supposed to check for. When receiving any type of input, you write checks for missing input, massive input, and malformed input. If you're not always doing those things, you shouldn't call yourself a tester (really the developer should have handled all these first anyway). This is one of the most basic QA things you're taught in any decent software engineering degree. Never trust input!

          I seem to recall browsers having issues with long URLs in the news a few years ago.

          If you read my post again, you will notice that I never said that Apple was blameless in this. I know that you should never trust input data. And I assure you that the iOS OS Dev. That wrote the code for that feature knows that, too. To assume otherwise is patently ridiculous, and you know it, or should...

          But what I pointed out was that I know from experience that it is an "understandable" error. You can get up on your high-horse all you want; but if you have been coding (or testing) for more than a year,

  • by 93 Escort Wagon ( 326346 ) on Tuesday March 29, 2016 @11:29AM (#51800123)

    I have DuckDuckGo set as my default, and I haven't seen this at all.

  • by Anonymous Coward

    It just works? or maybe not...

  • Thought it was me running out of storage.Happened with both iOS 9.2 and 9.3

  • This is why I never update iso. Updates are trash man. It's a portable sealed device, it's not like the hardware changed.
  • what is funny is that it appears the second link on this post is wrong too. from homepage returns 404 "glitch has something to do with Universal Links" => https://slashdot.org/techcrunc... [slashdot.org] from comments it appends the same link multiple times and opens the comments page https://apple.slashdot.org/sto... [slashdot.org] etc

"Necessity is the mother of invention" is a silly proverb. "Necessity is the mother of futile dodges" is much nearer the truth. -- Alfred North Whitehead

Working...