Apple Watch Bug Causing Restarts When Asking Siri About the Weather

[gnasher729]

How could this happen? This is so common and it does this.

I guess were weathering wrong?

As a developer: There is a software developer who created the bug, and a software tester who didn't find it. The "not finding the bug" is easily explained - it was (most likely) a bug that happened only on one day of the year, and if you test the phone on the 364 other days of the year, it works just fine. It's the kind of problem that is very, very hard to find.

How could that bug be created? That's hard to explain. Things can go wrong, obviously. But it should be very hard to take the watch down as it happened, even if there is a bug, so that is something that Apple should seriously investigate.

 

[MH01]

As a developer: There is a software developer who created the bug, and a software tester who didn't find it. The "not finding the bug" is easily explained - it was (most likely) a bug that happened only on one day of the year, and if you test the phone on the 364 other days of the year, it works just fine. It's the kind of problem that is very, very hard to find.

How could that bug be created? That's hard to explain. Things can go wrong, obviously. But it should be very hard to take the watch down as it happened, even if there is a bug, so that is something that Apple should seriously investigate.

Spot on . You can't find and fix all the bugs, though that is why you have betas, hoping people use the most common functions. Sounds like it was last code commits that were not fully regression tested. Should there be a regression test for the most common Siri requests ? Maybe - though guess not

 

[HarryWild]

Siri gets confused about the weather too! It sort of like for Siri, a person hanging up their phone when stared! Siri just to confuse to do any sort of reply! LOL!

 

[kingdomkids]

There's a strange bug with the Apple Watch today that's causing resprings whenever Siri is questioned about the weather. Asking Siri something like "What's the temperature?" or "What's the weather?" or "Is it raining?" causes the Apple Watch to crash.

The issue has been documented in several threads on the MacRumors forums and on reddit, and we've also been able to replicate it on our own devices.

Complaints about the problem appear to have started this morning, and the bug is confirmed to be affecting both LTE and GPS Apple Watch Series 3 models as well as older Apple Watch models running watchOS 4.1. Not all Apple Watch owners in all countries are affected, but it appears to be impacting users in the United States, Canada, and Europe.

It's not yet clear what's causing the issue, but the weather app itself is working fine, and restarting and resetting the watch don't appear to fix the problem. From MacRumors forum member SRLMJ23:Curiously, asking Siri about the weather tomorrow or next week doesn't cause a problem -- it's only questions about the current weather conditions that are resulting in errors. It's possible that this is an issue related to the upcoming time change, as one reddit user has discovered. Daylight Saving Time is set to end on Sunday, November 5 in many regions in the United States and Canada.Apple will likely have this fixed shortly (or it'll clear up after Sunday), but for now, it's best to avoid using the watch to ask questions about the weather in order to prevent sudden resprings.

(Thanks, Shawn!)

Article Link: Apple Watch Bug Causing Restarts When Asking Siri About the Weather

[doublepost=1509970611][/doublepost]It's not a bug. It shuts down so you will look up at the sky and feel the air around you to see what the weather is. Is supposed to be there.

 

[Tiger8]

Meh, there was a time Apple software was close to flawless, their quality control seems to get worse every year. This isn't the first time Apple software couldn't handle a daylight savings issue; remember alarm issues? You'd think they test for that

 

[BarracksSi]

“You’d think they test for that”... but I’m trying to figure out how they’d test a cellular-dependent device (yes, even the prior AW models) in an isolated, non-networked environment for this exact scenario, which has the national time servers also changing for DST.

I would guess that just reusing the same code as what it had before would be fine, but maybe there was enough stuff rewritten that they accidentally created a new conflict.

 

[maceric]

My Series 2 had no issues when I used Siri just now to ask for the weather.

 

[eric_n_dfw]

“You’d think they test for that”... but I’m trying to figure out how they’d test a cellular-dependent device (yes, even the prior AW models) in an isolated, non-networked environment for this exact scenario, which has the national time servers also changing for DST.

I would guess that just reusing the same code as what it had before would be fine, but maybe there was enough stuff rewritten that they accidentally created a new conflict.

DST change issues happen all the time in software. The "right" way to test for them has nothing to do with functional testing on the device itself but to add unit tests that use a "fake" (or mock as we say) clock and tests cases for boundary conditions like this as well as leap year related dates/times, etc. It's most likely that the developer of the weather app test code just never thought about it. (For non-developer types, unit-tests usually refer to very low level testing of individual functions/methods, not integrated systems/libraries/applications: https://en.wikipedia.org/wiki/Unit_testing )

 

[bruinsrme]

my watch did rhe reboot thing over the weekend. Tried it again today and no reboot

 

[kdarling]

DST change issues happen all the time in software. The "right" way to test for them has nothing to do with functional testing on the device itself but to add unit tests that use a "fake" (or mock as we say) clock and tests cases for boundary conditions like this as well as leap year related dates/times, etc.

This. After you're hit a few times with Daylight Savings and Year Change bugs, usually you look closer at your code. Been there, done that. Got woken up on Saturday mornings because of it

My guess is that someone subtracted some hours, didn't realize that for a short period of time the result would be -1, didn't check for that, and used it as a (non-existent) array index to write info to. Thus crushing some valid info, or causing an exception.

We had a project once where this happened way too often, so I actually changed the array allocation to include -1 as a valid index. Whenever we found something sitting in that spot, we knew a method somewhere wasn't checking. Often the data would give us a clue as to what did it.

 

[JorgTheElder]

During Apple's beta testing period, how exactly would they test for this?

Are you daft? You test it the same way you test all time zone/daylight-savings code. For part of the testing you set the device to different dates/times that are known to cause issues, like just before and after changes.

 

[kemal]

Not on mine (red dot.) But the other day it got hot while it was on my dangling wrist. Shut it down for a half hour. Fine now.

 

[Yojimbo007]

No problems on mine!

 

[69650]

Why didn’t they spot such an obvious bug in testing?

 

[noraa]

Are you daft? You test it the same way you test all time zone/daylight-savings code. For part of the testing you set the device to different dates/times that are known to cause issues, like just before and after changes.

Seriously, are you daft? You must have little to no conception on how iOS and Siri actually work. As such, let's think about this logically so you can understand how obscure this bug truly is. When you query Siri, Siri connects to Apple's servers. Apple's servers have their own date set, independent of whatever date is on your device (in this case your Apple Watch). As such, regardless of what date is set on the Watch, when you ask Siri what the weather is out TODAY (as in the actual today, not whatever date you set on your Apple Watch), Siri contacts Apple's servers, gets the real date, and provides you with the weather information. As such, it wouldn't matter what date is set on your Watch, as when you ask Siri about today - it will tell you about the real today, not whatever fictional today you set for your watch.

In other words, testing for this bug would only occur on ONE day out of the year, period. And Apple sure as hell isn't going to change the date on their servers to test for DST bugs, as that would screw up their entire customer base.

So yes, this is a very obscure and difficult bug to test for.

 

[bruinsrme]

Meh, there was a time Apple software was close to flawless, their quality control seems to get worse every year. This isn't the first time Apple software couldn't handle a daylight savings issue; remember alarm issues? You'd think they test for that

Is their quality getting worse or is the need to meet customer needs/expectation becoming more and more challenging.

Seriously, are you daft? You must have little to no conception on how iOS and Siri actually work. As such, let's think about this logically so you can understand how obscure this bug truly is. When you query Siri, Siri connects to Apple's servers. Apple's servers have their own date set, independent of whatever date is on your device (in this case your Apple Watch). As such, regardless of what date is set on the Watch, when you ask Siri what the weather is out TODAY (as in the actual today, not whatever date you set on your Apple Watch), Siri contacts Apple's servers, gets the real date, and provides you with the weather information. As such, it wouldn't matter what date is set on your Watch, as when you ask Siri about today - it will tell you about the real today, not whatever fictional today you set for your watch.

In other words, testing for this bug would only occur on ONE day out of the year, period. And Apple sure as hell isn't going to change the date on their servers to test for DST bugs, as that would screw up their entire customer base.

So yes, this is a very obscure and difficult bug to test for.

And yet we all survived this epic event as we did millions of years before the apple watch.

I thought it was pretty funny. When I tried it, the first time I thought it just shut down the watch. The second and thrid time it gave me the weather information then a second or two after it shutdown.

With the support of my family, friends, mac rumors and apple support I made it through these challenging times.

 

[witnessthefunk]

Seriously, are you daft? You must have little to no conception on how iOS and Siri actually work. As such, let's think about this logically so you can understand how obscure this bug truly is. When you query Siri, Siri connects to Apple's servers. Apple's servers have their own date set, independent of whatever date is on your device (in this case your Apple Watch). As such, regardless of what date is set on the Watch, when you ask Siri what the weather is out TODAY (as in the actual today, not whatever date you set on your Apple Watch), Siri contacts Apple's servers, gets the real date, and provides you with the weather information. As such, it wouldn't matter what date is set on your Watch, as when you ask Siri about today - it will tell you about the real today, not whatever fictional today you set for your watch.

In other words, testing for this bug would only occur on ONE day out of the year, period. And Apple sure as hell isn't going to change the date on their servers to test for DST bugs, as that would screw up their entire customer base.

So yes, this is a very obscure and difficult bug to test for.

I agree that this is a difficult bug to find but according to other users, this issue occurred in a similar way with another Apple product(s). When we test apps at the company I work for, we have multiple environments and test servers that we test on prior to moving code to prod. So although the bug is a difficult one to find, they should have the ability to run this test case without it affecting production.

 

[BarracksSi]

I agree that this is a difficult bug to find but according to other users, this issue occurred in a similar way with another Apple product(s). When we test apps at the company I work for, we have multiple environments and test servers that we test on prior to moving code to prod. So although the bug is a difficult one to find, they should have the ability to run this test case without it affecting production.

Yeah, and when it happened before, it was ONLY during the DST change — just like this one.

 

[witnessthefunk]

Yeah, and when it happened before, it was ONLY during the DST change — just like this one.

So are you suggesting that the only way to replicate this testing scenario would be on that one day? You do realize that DST is man made, therefore it can be replicated in a test environment. You may not be old enough to remember the “Y2K” scare but essentially companies were easily able to replicate that test scenario also. Nothing new here.

 

[Tiger8]

Is their quality getting worse or is the need to meet customer needs/expectation becoming more and more challenging.

It's both probably. But when customers pay $1000 for a phone (and potentially 10 times as much for a watch), you can hardly blame them for expecting Apple to test their code.