WHY can you not make Chrome update like every other sane, well-behaved app?
Update notification -> User confirmation (or an OPTION for auto-updating) -> Download status.
Why do you insist on installing things into our startup sequence without our permission? If your intent is to "protect" users, increase the nagging. I'd be fine with Chrome refusing to load any website until Chrome is updated to the latest version.
Even Apple, who is notorious for making users' decisions for them, lets us choose when to update apps and operating systems.
Google's software is a cascade of lies and deceptions.
Think about this: when you start to tamper with Keystone agent, it never says anything to you, it just silently reinstalls itself somewhere else like spyware.
It will keep asking over and over for root access, without explaining why. They make it seem like your installation is incomplete without root access, but that is a lie. It will function fine running out of ~/Library/ as /Library. But there is no way to make it stop asking.
Google Earth, Google Drive, or many other Google products will re-install Keystone agent.
If I try deleting it, then that means I probably want it gone. They should prompt me to repair it or leave it alone.
You would think that Google would want to show off their updater. Even just a growl notification that an update has occurred.
But it makes sense why they don't want users thinking about it. If they were more transparent, they would say: We've installed this software that will monitor your filesystem and make irreversible changes whenever we feel like it. Sometimes we will break things, but most of the time we won't and if we do break something, we will fix it. It is possible to disable, but you will have to search for it, because you will never discover it yourself. Oh, we could just have a checkbox in Preferences, but we want to make you work for it. And all you are doing is requesting that we stop updating, but we'll still be running.
They also used to shit on IE and WP, now everyone gets to enjoy Google's hegemony.
you have tons of complaints in this thread about google's "bad behavior" but you continue to put up with it to by patronizing the company and their tools, without even apparently asking the question, "do i really need chrome?" or whatever. have you asked yourself why you keep their software on your computer if it's such a headache?
i'm sure i'll get the typical "but there's nothing better!!" response and there may not be, but it's telling of you personally that you are willing to get so upset with all of this and then... keep on keeping on.
What would you suggest I do?
I use Chrome sometimes. Firefox is bad in its own way, often emulating the worst of Chrome. Like, at least the Keystone agent is unobtrusive and you don't even know it is there. Last time I checked, Firefox's Updater.app is just as disrespectful to the user, but it is horribly inefficient and clumsy.
Because someone at the Chrome team has setup an OKR of newest version rollout rate. His/Her salary and promotion are at stake.
Because that's how you end up with software that isn't updated, running old insecure versions.
As a user, I like it when my apps automatically update without me having to worry about it. The frustrating part about the Mac App Store is how it still makes you worry about updating apps.
Wait, what? The Mac App Store updates your apps automatically in background (I know bc sometimes it tells me it can’t update a particular app until I exit it)
that's the part I find annoying. Contrast to iOS which doesn't have this problem. Obviously the model on iOS is a lot different (more restrictive backgrounding, apps are build to handle shutdown at any time), but its still a minor frustration I have with MAS.
You see a download bar on app icons in the Dock and Finder while they are updating, then a badge (blue dot prefix before the name) on recently updated apps.
Rarely (i.e. on new user accounts) it may ask you for the iCloud account (if it was a purchased app, I think) or administrator password (after some major OS installations).
How is that frustrating and "making you worry" about updating?
It seems like you haven't used the Mac App Store or have changed the default to manual updates.
Obviously the model here is different, but its still a minor frustration to me.
There is no legitimate reason for any install other than an OS upgrade to modify /var or any other system-related directory.
Ever.
> We're building a replacement that fixes the problem.
The fact that your team would allow any code which modifies a machine at the OS-level only reifies the concerns regarding Google's products.
rm -rf /Library/Google/GoogleSoftwareUpdate/GoogleSoftwareUpdate.bundle
ln -shf /private/var /var
chflags -h restricted /var
chflags -h hidden /var
xattr -sw com.apple.rootless "" /varTell your team and your supervisors this:
My computer is not your playground.
There is no legitimate reason for a user-space install to manipulate system directories. So for an install to do so, there must have been an conscious decision made and code written to make real.
Therefore, for this system manipulation to have both been introduced and released, "most engineers on the team" either raised no problems with it or did not consider the implications of this decision.
> But for the time being, let’s let engineers do their job?
They did their job, which resulted in the release of this system destabilizing product.
Perhaps the job they should have done was to consider their work product be one which did not assume complete control of the machine onto which it runs?
Users have no choice but to take whatever updates they throw at us, and have no recourse but to sit around and wait for another update to be pushed.
There is no way to roll updates back, and disabling updates is obfuscated and hidden away behind an obscure terminal command that nobody would discover on their own.
Google invited themselves into the guts of our computer on the pretense of updating their browser, and then they made a mess.
If Google explicitly laid out what they were doing and asked permission, many users would not grant it, which is why they are so covert about it. It isn't that it is being unobtrusive, it is that it is hiding.
I swear, only Google can get away with this. Nobody was this defensive when Microsoft pushed Windows 10 on people.
These devs know what they are doing and in the current economic environment are clearly happy to be doing it.
Killing it resulted in a relaunch and 100% cpu. There is no way to stop this except for unloading the launch agent, AND launchdaemon. Removing the application and killing the instance.
The os platform providers updates.. use that instead of crafting your own malware.
How would you like it if your car suddenly has a top speed of 15mph, and no power steering, because someone wanted to update the number of radio presets.
Of course then it stopped booting at all yesterday so if it was that then it must have pushed the 4.5yr old battery over the edge and killed it. Or overheated something until it died. I don't think those fans have ever been cleaned.
Makes me wonder if other software might be attempting to damage the system (totally by mistake) but SIP is preventing it, making it quite deadly to use said good software if you happen to turn off SIP for stuff like debugging
Seems an awful lot of work related computers (eg Avid systems, and more) have been rendered inoperatable until someone manually boots and fixes each one.
touch ~/Library/Google/GoogleSoftwareUpdate
touch /Library/Google/GoogleSoftwareUpdate
chmod 000 ~/Library/Google/GoogleSoftwareUpdate
chmod 000 /Library/Google/GoogleSoftwareUpdateIt seems kind of counter productive to kill off the auto update system when you can just as easily switch to a browser that just doesn't do what Keystone does.
More important, I like Google Earth and it tries to install Keystone too.
I've heard "Firefox is better than it was" only for me to reinstall the latest and find it's still way cludgier than chrome.
In general, I've found it to be much better than Chrome, but as always YMMV.
I think in the end that's something that you have to test out for yourself periodically, as it seems to be great differences of which is the best performer across OSs and devices. As a rule of thumb I try to do a short evaluation of each of them every ~5 releases.
From experience, they absolutely do it.
Thank you. This was the best tip to come out of this whole discussion.
I'll identify every location Google apps write to, and lock them out with this.
They are both very low, but I’d rather programs not change my stuff against my will, even if they are trying to protect me.
Problem partly solved.
My wife's a primary school headteacher (or K-12 as you say in the States). Her MacBook was disabled by this. Yes, she takes weekly backups, but schools don't have free money to spend on spare laptops for a few days' work, nor on unnecessary technician time to fix it. Fortunately I spotted this posting (thanks, HN poster!) on blearily checking HN this morning and instantly recognised this was what's happening.
Have some decency for the people whose lives you've just affected and apologise to them.
Your true statement that tone will often matter is an interesting discusson on society and education. That it is also relevent on a site otherwise dedicated to intelligent discourse was the nugget I was hoping people would think about.
Couldn't help but notice most of the posts from the Chrome team lacked any sort of apology. Including this support post https://support.google.com/chrome/thread/15235262
Bit disappointing when this bug has bricked multiple machines.
> If you have not taken steps to disable System Integrity Protection and your computer is on OS X 10.9 or later, this issue cannot affect you.
SIP came in with 10.11. Any machine on 10.9 or 10.10 is vulnerable.
Do you want only PR people on HN trying to talk to you? Because this is how that happens.
One of the first places "I'd" look? It's not my Mac. I'm not sure how many primary headteachers read Hacker News or have a spouse who does. I'm guessing <1%.
When the world's biggest software company actually bricks people's Macs with a software update, then "sorry" is the least I expect, frankly. But if you want to dismiss this with "dickhead", you do you.
I'm torn on that one. I want direct communication to be possible without running it though PR or people with PR training, to improve response times especially in such "busy" situation. This requires us on the receiving end to be somewhat lenient. But on the other hand, I also don't find something better elsewhere, including the more official announcement[0] linked to. Thus this style seems like company policy and certainly deserves criticism.
The keystone team accepted this as a p0 bug.
/ also shouldn’t be writable so it’s understandable how this wouldn’t have been caught. writable / is not sane and it’s unlikely you’d test that case.
The version of Keystone packaged with Chrome is not affected by this bug, so allowing it to reinstall Keystone will not recreate the issue.