Microsoft Has Broken Millions of Webcams with Windows 10 Anniversary Update
thurrott.com
thurrott.com
In addition the update cleared a bunch of my settings, including silently switching on some of the privacy-questionable features I'd opted out of previously. The update process itself wasn't pleasant, either: while I was away from my desk my laptop woke up from sleep, threw away everything I had open (including a running VM instance, risking data corruption), and launched into a patching process that cost me nearly an hour of work time.
I understand the importance of automatic updates security-wise, but Microsoft seems to be actively user-hostile here.
</rant>
It also added Edge to my taskbar but did not replace either Firefox or Chrome. Are you sure it replaced it?
Seems like some misinformation here.
You can create a desktop icon with a link to IE/Edge, Firefox, Chrome, but you can create also a desktop icon with a link to the current default browser.
The problem is that if you have selected Firefox as your default browser, then one of the default icons looks like a link to Firefox, with the Firefox icon, but it actually a link to the default browser. If you copy or drag&drop it, then it look like Firefox, but it's a link to the default browser.
It's very difficult to spot the difference, and after a few months you will forget what kind of link it was.
If the Windows Update changes the default browser (perhaps automatically, perhaps a confusing question in a popup), then your icon to Firefox will change to an icon to Edge.
(I had a similar problem when I changed the default browser form IE to Chrome last year in one of the computers at work.)
I did already have the Edge icon on my taskbar (pre-update) but the Chrome icon remained, as did it the default.
* Turned Fastboot back on, breaking my Linux dual-boot.
* Enabled remote desktop.
* All privacy settings set back to default.
All of these settings were previously turned off by me. I'm now in the process of removing MS products from my computers.
I install Linux (usually Lubuntu or Ubuntu once or twice Mint and Arch) on my dev machine about once a year.
In a few days (my current record is a week, last attempt <24hrs) I uninstall it.
Nothing just works. I was happy Wifi was working last try, then my computer crashed coming out of sleep. My hi resolution monitor doesn't work properly with a standard monitor. A few tries ago I couldn't even boot after install because of my NVIDIA card.
There's no standard for anything. On OSX setting my shell via chsh worked, Ubuntu ignores the same exact command. Applications can't just register themselves without distro specific tweaks so half don't bother, so every install I was firing up alacarte
For all the trouble I get a "free" OS, but do I need one? Do I really care if MS is tracking my clicks if it means I don't have to deal with broken drivers? Do I really care if RDP is enabled when half the time I was fixing stuff I was downloading source code I didn't take time to understand to compile and execute with full privileges?
The rabid hate people who write software for a living have for closed source software makes me feel like I'm doing something wrong. Do these people just have super human capabilities to be able to keep up Linux dev machines and resist the temptation of working MS installs? Is Windows as bad as people claim and I'm just subject to Stockholm Syndrome?
Am I doing something wrong if all three of those things happening to me wouldn't be enough to even dream of installing Linux again? Because I feel Linux has done worst things to me for less benefit?
The only issue you discuss that I've experienced is the problem with mixed-DPI monitors, and it's supposed to be resolved in the next release of GNOME.
I don't hate closed source software (Plex is probably my favorite software product ever), but I certainly have an extreme distaste for Windows. I've wasted way more time reinstalling Windows than I have debugging any Linux installs. (I'm at ~4.5 years of rolling releases in Arch and I've only had to take corrective action once or twice and it was announced ahead of time. And I'm running [testing] repos...)
When I read things like "On OSX setting my shell via chsh worked, Ubuntu ignores the same exact command" I start making some assumptions about user error since `chsh` works just fine in Ubuntu...
And then this "Applications can't just register themselves without distro specific tweaks so half don't bother, so every install I was firing up alacarte" simply hasn't been true for... 8+ years. I literally had to look up alacarte to remember what it is. Alacarte doesn't even do anything in a GNOME3 environment.
But anyway... I don't think anyone should really feel guilty about their OS preference. (I also think that most of the anti-Windows10 FUD is completely blown out of proportion and often borders on dishonest.)
And the shell problem is just me describing a situation exactly as it happened, chsh would not change the terminal that Gnome Terminal used, I had to manually change another setting. It's not that chsh didn't work, it's that the default terminal application installed in the DE didn't respect it. OSX's default terminal respected it just fine. It's not a major gripe, I just thought it was suboptimal.
And on a somewhat unrelated note, I've always noticed if I say something is broken in OSX or Windows I get the usual "Well what do you expect, it's OSX/Windows" (usually with a "What are you doing in X install `whatispetdistro`"). With Linux I get "Well I don't know what you're doing wrong because X has worked for me just fine". Half the time what they considered fine is what I consider broken or suboptimal. But that's anecdotal, it might just be the people I've interacted with in the past.
Alright, I'm curious... what software and what DE? The XDG spec defines a standard place to look for these files, and it's honored by KDE3, KDE4, Gnome2, Gnome3, MATE, Cinnamon, Budgie, ... and probably more that I'm forgetting.
So, if it's not happening, it's either a bug on the software for flat out not shipping it, or a packaging bug on the distribution's part for not placing in the correct place. But like I said, I've not encountered this in the wild for longer than I can remember.
>And the shell problem is just me describing a situation exactly as it happened, chsh would not change the terminal that Gnome Terminal used, I had to manually change another setting.
:/ I just tried it on a 16.04 install and on an Arch Linux install it worked as expected. Gnome-terminal defaults to "Run my shell" but has options to override it which are not on by default. You can see here where it uses your user set shell: http://bazaar.launchpad.net/~vcs-imports/gnome-terminal/mast...
>With Linux I get "Well I don't know what you're doing wrong because X has worked for me just fine".
From my experience, most times when I sit down with a user, and walk through troubleshooting, I see things, ask and hear back "Oh right, I forgot I tried fiddling with that too, did that make it worse, oops!".
If these things are so easily repro'd, there are people who would love to have bug reports for them. I've filed a couple GNOME bugs before, it was a straightforward experience.
>no standard for anything there is a standard, it's just not the standard you're "used" to. again, if you don't wanna spend time learning new paradigms-linux is not the way to go for you
>The rabid hate people who write software for a living have for closed source software makes me feel like I'm doing something wrong. Do these people just have super human capabilities to be able to keep up Linux dev machines and resist the temptation of working MS installs? Is Windows as bad as people claim and I'm just subject to Stockholm Syndrome? probably, linux requires a "i'll fix it/patch it/install something additional if it doesn't work" mindset whereas MS and Apple sell a "it just works" mindset to you-which works until something breaks(eg windows 10 broke my wifi,had to do some drilling to figure out windows 10 was using some kind of wonky setting for it that my wifi chip didn't support)
I'm not sure how I think about this. On the one hand it's a bit sneaky, on the other hand it was pretty clearly written (click here to try a new browser, click here to continue using Firefox), and I don't know how you'd get somebody to try a new browser otherwise. Google hat at one point agressive Chrome ads on their main page. And Edge is actually a surprizingly decent browser, so if they ask me once nicely, and don't nag, I don't have a big problem with it.
I'm in the EU btw., I'm sure MS uses different approaches in different places. Replacing a pinned icon is really rude, especially since MS said this area belongs to the user, and no developer, not even MS, is supposed to change anything there.
https://social.msdn.microsoft.com/Forums/windowsdesktop/en-U...
That's true though: why is it important for them?
To think, Microsoft used to be all about preserving backwards compatibility to the extreme.
No problem: I found lots of solutions online. Problem: none worked, and the ultimate suggestion was to refresh or reset. Refresh didn't work, so I had to reset losing all of my installed apps and settings.
When given the choice, I just went back to Win 8.1. I am astounded that a 'version 10' of any software would have such a keystone issue, at least losing the ability to navigate my GUI without crazy right clicks and searches.
I run Mac and Linux too, so I am not a fanboy of any one particular OS. I call problems out on all of them equally. I just want them not to bother me while I am working ;) Or waste a day reinstalling apps and re-registering them, and setting them.
In 38 years of owning and running computers, two systems stand out as particularly robust - my 1978 Commodore PET with built-in cassette tape drive storage, and a FreeBSD box I was running for a long time until the hard drive failed due to operator error!
So basically I have to keep a secondary PC around in case my Windows 10 PC ever reboots, so I can have a mouse and keyboard. I saw lots of posts online about keyboard and mouse not working on boot in Windows 10, but unplugging/replugging seemed to work for those people. It did nothing for me. I saw advice to disable fast boot, but even after doing that, on restart they stopped working. It also seems pretty suspect that the keyboard and mouse work fine in my BIOS and only stop responding after the Windows 10 spinner appears.
I mean, if they're serious about this, making Windows open source would make sense, so that external testers can also contribute patches for consideration.
In case your genuine was genuine, I'll give a somewhat serious answer: The intention was that the SDET work would be picked up by SWEs who would be responsible with dev and test. Ops also is slowly being absorbed in this manner (this current wonderful permutation of the word "devops")
As to my thoughts on the transition... let the assertion that my home machines are still (happily) running 7 speak for itself; and I say that with unfortunate tone in that I'm a big fan of hyper-v integration/DX12.
My thought on Windows >7 is that 8 and 10 have introduced many desirable improvements in various subsystems, but unfortunately one cannot have the old user experience (classic shell) with, say, new core subsystems and core system apps. Thankfully, I don't depend on Windows, so it doesn't affect me directly, though the times I have interacted with Windows 10 on a friend's laptop, I couldn't believe how much it regressed for power users. Simple things like login and logout are totally messed up now, as is the waste of screen real estate and Windows 1.0 style widgets with a flat theme plus low-contrast color palette, large-white regions. To top it off, even if one creates a custom install image with tools like nLite, Microsoft is known to reinstall stuff one absolutely does not want. Finally, the mobile platform inspired data collection services are out of place on the desktop, unless they're opt-in and optionally installed.
I once tried to use a Windows 8 tablet when it was new, and I could not figure out how to use the touch interface and there were no indicators, so I concluded that there must be a getting started tutorial one sees on first boot of the device. KDE and GNOME are repeating the same mistakes, by trying to make the desktop something it isn't. It's like trying to turn a PC into an Xbox with a WiiU controller, taking away precise input devices and efficient use of screen space. Making widgets large to suit touch use makes no sense if you control it with a keyboard and mouse.
Internet sources didn't give a clear answer about what's being sent, but some sources seem to think it's full keylogging. If that's true, Microsoft is in for litigation and I'll be first in line to sue.
Thurrott's explanations is incorrect. What is happening is that Windows is decoding the MJPEG or H.264 encoded streams coming from certain high end webcams and delivering the decoded frames to the video chat application[0].
[0] - https://social.msdn.microsoft.com/Forums/windowsdesktop/en-U...
Even if you have a secondary storage (not external), it refuses to use it. Tried looking on the net, and microsoft official answer is "nope".
It's like they really, really don't want to be on that segment, despite trying really hard at the same time.
Apparently, you're not the only one with that problem...
Despite being amusing to drive-by my wife whilst she was watching cat videos on Facebook, it did make me acutely aware that testing every conceivable hardware combo is a big ask, even for MS.
Still, this seems like something they absolutely should have picked up with even rudimentary QA - millions of users represents a non-trivial profile, and they would surely have been aware of the update.
I'd assume the Skype QA cycle is focused on new releases of their own codebase rather than an ongoing maintenance cycle? The take-away here is MS need to internally flag potentially breaking changes more thoroughly.
Source: https://news.ycombinator.com/item?id=12313179
So it seems that it not only Linux, but also Mac. So much that it's making every desperate move to make users switch over to these alternates.
Well done!
- Sent from my openSUSE Tumbleweed installation
It is a shame that Windows has no way for me to find out which program is actually using the webcam so that I can terminate it. Unlike Linux of course!
In 6 months, I had to restore windows 10 3 times after botched updates.
Having said that, I've performed an uncounted greater number of upgrades than I can remember and have only OCCASIONALLY hit this issue, and it was always when I was doing something outside of the distribution recommended set of instructions (more akin to trying to go from WinXP directly to Win10 without any intermediate upgrades).
There's plenty examples out there of Linux kernel regressions that have been introduced along the route, there's even a full paper on tracking them. Then there's plenty of examples of things breaking on upgrades in Debian/Ubuntu/Fedora/CentOS/Arch, systemd caused some trouble too and so on and so forth, and not just on major boundaries.
I was simply curious why the parent commenter had an axe to grind with Linux. Everything has its bugs.
I have a laptop with integrated Intel GPU. Not powerful but good enough to play a few simple games from time to time. After updating from Ubuntu 14.04 to 16.04 most games won't work properly anymore: they run fine for a while but then the display freezes for a few seconds. Most of the time after that is start working again, some other times the game will crash (I am blaming the GPU since sound and games logic continue regularly, and CPU usage is unaffected).
I have also a raspberry pi. One day, after a routine update the sound started behaving weirdly, sometimes going faster/slower, some other times burping weird sounds. As with most audio problems in Linux the solution was simply to nuke pulseaudio from the system.
Unless you're making an independent film or something, I would think that a high end webcam does a great job for professional streamers and online meetings. Even the cameras in modern laptops and smartphones do a good enough job.
how much QA should they do to verify the millions of different hardware and software combinations out there won't have some edge case behavior?
You want them to track your privacy settings so they can be applied. Oh but then they are tracking us!
Cause everyone here works at a company that gets their code right every time, I'm sure
The privacy settings do not need to be transmitted back to Microsoft. At the start of the upgrade, the installer could read the current settings, then reapply those once the install was complete. No privacy-infringing tracking required.
This argument doesn't hold up. Privacy settings can be stored locally, which requires no tracking. The problem arises when those privacy settings aren't correctly applied or remembered between updates.
As much as they used to do. They cut down on QA after Windows 7 (see laying off test engineers and laying the burden onto the primary development engineers)