Please create an account to participate in the Slashdot moderation system

 


Forgot your password?
Close
typodupeerror
×
Bug Apple

Some Apple Watch Series 4 Models Are Frequently Crashing and Rebooting Due to a Daylight Saving Time Bug (macrumors.com) 110

Posted by msmash from the how-hard-could-it-be dept.
Some Apple Watch Series 4 owners in Australia experienced crashes and reboots on Saturday due to a bug that surfaced because of the daylight saving time change. From a report: According to Reddit users hit by the Apple Watch bug, the root of the problem appears to be the Infograph Modular face's Activity complication, which displays a timeline graph with hourly data for the user's Move calories, Exercise minutes, and Stand hours. When daylight saving time (DST) lops an hour off the typical 24-hour day, the Activity complication is apparently unable to compute the change and draw the timeline graph with only 23 hours, which throws the Apple Watch into an endless reboot loop until the battery runs out.
This discussion has been archived. No new comments can be posted.

Some Apple Watch Series 4 Models Are Frequently Crashing and Rebooting Due to a Daylight Saving Time Bug More

Some Apple Watch Series 4 Models Are Frequently Crashing and Rebooting Due to a Daylight Saving Time Bug

Comments Filter:

  • Whoopsie (Score:4, Insightful)

    by war4peace ( 1628283 ) on Monday October 08, 2018 @08:07AM (#57444454)

    One more reason to do away with that monstrosity called DST”.

    • Re:Whoopsie (Score:4, Funny)

      by phantomfive ( 622387 ) on Monday October 08, 2018 @08:36AM (#57444632) Journal
      No, this is a reason to keep DST. So we can easily see and laugh at the incompetent programmers.

      • Re: (Score:3, Insightful)

        by angel'o'sphere ( 80593 )

        Well, from one point of view you are right.

        However, as I worked quite a long time in the energy business: DST is a pain in the ass. Not only because of programmers making mistakes (I would not call that incompetent ... what again are the rules for a leap year? Are you sure you know them all? It is only 2 rules, though, or 3, depending how you count)

        Mistakes are also made in the requirements/business analysis. And if a programmer says: "uh, that does not look right" often he gets stared down and challenged:

        • OK, that's fine, but if daylight savings time throws your os into an endless reboot loop something is wrong with your head.

        • One unix idiot ... idiot because I considered him a friend, until he started bullying me in some jobs ... once said: "it is so easy! Just schedule everything in UT!" What he did not grasp is: everything is scheduled in UTC. But you nevertheless want one day in the year to see your whole day as a 25h day, and the other day in the year you want to see your whole day as a 23h day ...

          What makes time extra difficult is that so many people assume that time is easy. And because people mistakenly assume it is easy, even very very smart people whom you would probably say are reasonably careful will get it wrong surprisingly often. Having a big brain and making a good guess is not good enough here.

          • What makes time extra difficult is that so many people assume that time is easy.
            Exactly!

            And because people mistakenly assume it is easy, even very very smart people whom you would probably say are reasonably careful will get it wrong surprisingly often.
            Yeah, somewhere is a web site with all the "misconceptions about time", they also have some other topics, like the misconceptions of "Names of Humans".

            Ah, found them:
            https://www.kalzumeus.com/2010... [kalzumeus.com]
            https://infiniteundo.com/post/... [infiniteundo.com]

            • Thank you, those are awesome.

              I like this one: "People have names."
              I would find it hard to believe that anyone has no name at all, but I can easily imagine everyone in the immediate family disagreeing what the name is.

              Apparently in the Middle East, it is common to refer to someone by personal relation. So in one context I might be "Bin Muhammad" which means "We all know that Muhammed guy (not to be confused with the other 9 Muhammeds we mutually know), and this is his son". And in another context, the same

              • One confusing part about names are: they can change, the simplest case is marriage.

                Now I worked on a software for planning and organizing schools. Classes, students/pupils and teachers.
                A school year consists of classes like 5a, 5b, 5c and 6a, 6b, 6c and so on. Depending on the grade as in 5th versus 6th grade, a class has a set of courses/topics. Obviously in Germany minimum 2 different religious courses, often more (as islam and judaism is now mandatory in schools for students of that religion, years ago o

        • I would not call that incompetent ... what again are the rules for a leap year? Are you sure you know them all? It is only 2 rules, though, or 3, depending how you count

          I agree it's not incompetent. Incompetent is an understatement; it's retarded.

          I knew the rules when I was about 10.

          • So did I,

            but I would wager 95% of the people on the planet don't.

            Now we could argue that most software developers belong to the upper range of IQ and skills, but I still would bet 50% of all software developers don't know the rules.

        • The snag is that the real world wants local time. Introducing UTC makes some things siimpler, but it also means you need to be converting time for some operations. For example, there are real world business needs for knowing what the daily power usage was starting at 12:00AM local time, they don't really care about what this means in a neighboring state or country. Meanwhile there may be an operations manager panicking because he thinks billing software will screw up if they feed in 25 hours in a single

          • Meanwhile there may be an operations manager panicking because he thinks billing software will screw up if they feed in 25 hours in a single day.
            That is one of the points why "time problems" are not trivial, even if they look like that.

            In my experience (in software engineering) basically everything that looks trivial on the first glance gets pretty quickly quite complex.

            And migrating away from time zones and DST is amazingly hard to do, even if some programmers think otherwise.
            Exactly.

            I first stumbled over

    • Re:Whoopsie (Score:5, Insightful)

      by ledow ( 319597 ) on Monday October 08, 2018 @08:46AM (#57444690) Homepage

      But you'd rather these people who can't use long-established timezone libraries, handling protocols and data formats did things like monitor your heart rate and ran your car instead?

      The problem here is not DST or the complexity of it... if that were the case the banking industry, Internet, phones, etc. would fall over at the same time because of the same problems.

      The problem is that someone at Apple doesn't understand how to handle dates properly despite there being long-established libraries for exactly that. And they didn't bother to check (i.e. test) adequately enough before releasing millions of dollars of mass-market products.

      The problem is lax development and testing. Not the complexity of handling something that billions of devices owned by billions of people around the world handle every year just fine.

      • No, I understand that. In no case was I making a point to save Apple's incompetence. It's just that DST shouldn't exist anymore.

      • if that were the case the banking industry, Internet, phones, etc. would fall over at the same time because of the same problems.
        I wold dare to say: at least the banking industry is completely uneffeced by DST, and I'm pretty sure phone companies and internet companies are neither.

        And they didn't bother to check (i.e. test) adequately enough before releasing millions of dollars of mass-market products.
        Checking/testing for time bugs is a pain in the ass. If you have a solution you are a millionaire over nig

    • One more reason to do away with that monstrosity called DST”.

      Hear, Hear!

    • It's a problem to be sure, and I've stumbled into it several times in the past at different companies. Logic says to just use UTC and simplify it all, then convert to local time only when necessary to display the time. But in reality, the customers and product managers want local time and they don't really understand that some days have 23 hours and some have 25.

      Twice a year at one job I had to write up a summary of what would happen with data collection that straddled a time change. Every single time it

  • It just works...mostly

    • Re: (Score:1)

      by AmiMoJo ( 196126 )

      Apple seems to have a problem handling time. Recall that they had problems with alarms not working a few times near the start of the year, and a soft brick issue if the clock was reset to 1970.

      I'd love to know where these problems are coming from. Are they using their own internal time handling library code? Because these issues are fairly unique to Apple, they don't affect the underlying BSD operating system that iOS is built on or any of the open source libraries preferred by Google.

  • Why now? (Score:2, Informative)

    by Not-a-Neg ( 743469 )

    DST doesn't end until November 4th.

  • Apple Watch Series 4 was introduced in September this year. DST doesn't change until November.

    How is this already a problem? Testing?

    Oh, wait. Apple doesn't actually test stuff, but somewhere in the rest of the world someone does...

    • What is really bizarre is that Apple seems to see no end to humiliating DST/time-based bugs. You would think they would have standard libraries/development guides/mandatory test cases related to this stuff by now.

    • DST in Europe ends last WE of October ... just in case you are wondering where the bugs come from.

      Uh, and if you don't know how software is tested I would refrain from stupid statements like "Oh, wait. Apple doesn't actually test stuff, but somewhere in the rest of the world someone does..."

      Last time I made an omelet, I burned it on one side. How would testing prevent me from burning it? It does not. The only test I can make is: put it on the plate for the customer and realize: oh, it is burned.

      For time rel

      • Apparently DST ends in Australia this past weekend. Three different change dates I know now, thanks gang.

        Um, I do actually know how software can be tested, but forgot the /s flag for you.

        And just for future reference, when I burn an omelette I often realize it before it comes out of the pan. But that's just me. I actually flip it once more before I slide it out of the pan. Easy test, and the plate is avoided.

        Oh, and if GPS time is such a problem, oh, wait, what? /sarcasm.

        • Re: (Score:2)

          by jrumney ( 197329 )

          Apparently DST ends in Australia this past weekend.

          Ahem. It began this past weekend. On devices that support timezones properly, it will end in April next year. On other devices, it may end on Dec 31, or it may just have failed to begin, as date > dst.end_date == true already.

          And woe to you if you live further East in New Zealand, where the impossibility of being 13 hours ahead of GMT just began.

      • Yep, you need ability to test the device; which means ability to turn off time synching with NTP or servers or a way to use special testing servers, stick it on a closed network, etc. I have had QA tell me "but if I change time on the device it will screw up all my other tests!", but that's just how it is so schedule extra testing time.

    • DST changes at different times in different countries. The November change in the US is particular to a relatively recent act of congress; many countries follow the US lead but also many do not.

  • The same thing happens to Linux using Oracle VM. You have to run sudo ntpdate -s [ntp server] or ntpd -q when using Linux virtualization sometimes because there is clock skew or an internal error. Until one changes the date to the current time on their Linux machine they cannot browse the Web because of security certificate issues complaining about the wrong date.
  • Indiana!

  • I knew it! (Score:4, Interesting)

    by TeknoHog ( 164938 ) on Monday October 08, 2018 @12:42PM (#57446228) Homepage Journal

    Daylight 'saving' time is a bug in any sensible timekeeping system.

    This is a timely (ahem) topic in the EU, as we have recently decided to ditch the biannual idiocy for fixed timezones. The remaining problem is that member countries can decide on which timezone to use, which might not be the solar one. So proponents of DST are campaigning for year-round DST (solar + 1). To me that sounds even sillier than current DST changes -- if you're going to make it permanent, you might as well fix your working schedule instead of redefining time itself. ("I want to go to work 1 hour earlier, but I still want it to be 8 o'clock, so I'll just move the definition of 8 around until I'm satisfied.")

    Meanwhile, several European countries are already on a non-solar timezone in favour of Central European Time. It's an interestingly rational alternative to solar time, but for us Finns it would be the "solar - 1" zone, in direct opposition to the DST camp. So one can hope that we'll compromise on the solar zone.

    • This is a timely (ahem) topic in the EU, as we have recently decided to ditch the biannual idiocy for fixed timezones.

      Well somebody proposed it. Nearly the same.

    • It still leaves the bugs of having to support multiple time zones and local variants.

    • you might as well fix your working schedule instead of redefining time itself

      This is a great idea. I suggest we pick a standardised metric around which we can align a widely different group of people, businesses, schools, and society as a whole. To make it easy we could divide it into segments called "hours" and so people remember how it works we could make the top most hour roughly the time when the sun is in the sky.

      This may just work!

      I want to go to work 1 hour earlier

      So do I but then who will take the kids to school?

      So one can hope that we'll compromise on the solar zone.

      Why? What's this obsession with aligning time so the sun is in the peak at noon rather than extend

      • What's this obsession with aligning time so the sun is in the peak at noon rather than extending the usable light at the end of the work day?

        You answered your own question. If you're interested in usable light, then it's nice to know where the sun is in its daily cycle. It's easier when the peak sun is an "origin" or "zero point" in the system of measures, not some arbitrary number.

        • No I did not. The sun at its peak in the sky at noon is not usable light for the vast majority of people given the asymetric work day and the general approach to life that prioritises after work social life over anything that happens before work.

          So I ask you again, I want to maximise my usable time, why are you obsessed with the sun being at its high point at noon rather than being at its high point at 2pm which would suit the needs of society better? Noon is just as arbitrary as any other point in time.

          Unl

          • why are you obsessed with the sun being at its high point at noon rather than being at its high point at 2pm which would suit the needs of society better? Noon is just as arbitrary as any other point in time.

            I tried to explain this with my 8 o'clock example earlier, so let me elaborate. Here the standard workday is 8 am to 4 pm, symmetric about the clock noon. But people wanted more light time after work, so they came up with DST. Now they still work 8am-4pm, but solar noon happens at 1 pm. My question is, couldn't they have shifted working hours into 7am-3pm instead? Because something has been shifted -- the relation between the Sun's cycle and their working hours. Of course, you get the same relational effec

            • My question is, couldn't they have shifted working hours into 7am-3pm instead?

              Since we're going around in circles, sure some people could have shifted. Will the schools, the banks, the subcontractors, the delivery companies also shift? Or maybe it's just simpler to change the timezone, not have to have the entire world re schedule re adjust, and just face the fact that there is no technical or social reason that the day should be symetrical around a high sun at noon*.

              *And where I live it isn't. The sun is at its peak around 11 or 10 depending on DST.

  • It's a Daylight Savings Time feature.

    Stop using Daylight Savings Time.

    Problem solved.

  • You would think with the endless string of daylight savings time related problems that Apple has faced in the past they'd have a dedicated part of the QC now checking to see how every device operates at the day DST changes.

Slashdot Top Deals

No spitting on the Bus! Thank you, The Mgt.

Close