an upgrade to macos 12.0.1 monterey, and running a simple privacy test on four browsers afterwards

Yesterday I finally upgraded my MacBook Pro (mid-2019, i9, 64GiB memory, 4TB SSD) to macOS 12.0.1 Monterey. It’s not like I appeared to have a choice. In attempting to upgrade to macOS 11.6.1 I triggered the further update process. I certainly didn’t want to, not after reading that Monterey 12.0 bricked a number of MacBook Pros, some Intel and some Apple Silicon. I’d wanted to wait until Monterey 12.1, but such was not to be. So far the only problem is that my fix for allowing VirtualBox virtual machines to run has been reset, so I’m going to have to go through that process as documented for Big Sur ( https://arcanesciencelab.wordpress.com/2021/10/30/running-oracle-virtualbox-on-macos-11-6-big-sur/ ) once again.

I then checked the four browsers I have on this machine by running EFF’s Cover Your Tracks privacy test ( https://coveryourtracks.eff.org/ ) just to see what happened. The browsers I tested (listing extensions where used) are:

  • Chrome Version 95.0.4638.69 (Official Build) (x86_64)
    • Using Adblock Plus and Privacy Badger
  • Vivaldi 4.3.2439.65 (Stable channel) (x86_64)
    • Using Javascript Restrictor
  • Firefox 94.0.1 (64-bit)
    • Using Privacy Badger
  • Safari Version 15.1 (17612.2.9.1.20)

They’ll be presented in that order below.

Everybody loves to hate on Google, especially Google Chrome. I’ve read more than one article on the web where the author hysterically exhorts their readers to delete Chrome before it sucks the end user’s very soul into the black depths of Alphabet’s digital hell. I suppose if you’re that paranoid then by all means drop Chrome. Except EFF’s test gave Chrome a very good score with regards to tracking, perhaps the highest a browser can get. I’ll keep Chrome around until there’s a real need to move on. And an interesting note: when I run the same test on my Chromebook running the latest Chrome OS, the score is the same as here.

Vivaldi was the second browser I tested. The first time I ran the test I was shocked when it came back as NOT blocking tracking ads and invisible trackers. Paradoxically fingerprinting was randomized. The shock came because months back I’d run the same test and blocking was in place. This meant that something had changed in Vivaldi, perhaps due to something I’d inadvertantly done. So I went exploring and found the following:

You’re looking at Vivaldi’s Preferences, Privacy section. The Blocking Level is towards the middle of the page. When I first opened it the level was set to No Blocking. I set it to Block Trackers and Ads and re-ran the test. The test came back with satisfactory blocking (yes in both columns as seen above). I don’t know how that changed, because I’ve never touched that except for the first time I set it to what you see now. The only way I know it could have changed is through an update. Lesson learned is to go back and check this again on every update in the future, which is somewhat sad as I’ve never had a browser flip a setting on an update.

The third browser I tested was Firefox. Note that the fingerprinting is unique, unlike Chrome and Vivaldi, who have random fingerprinting. Also note that the tabs across the top are the button-like tabs, which I don’t particularly care for. I’ve reached a point where I tend not to use Firefox, even on Linux.

Finally we come to Safari. It has the worst score of the four browsers I tested, even though Apple makes loud pledges about privacy and security. Safari on macOS (and probably everywhere else it runs) gives one pause to wonder if Apple is all talk in this area. I think they are.

I also would like to point out that the tabs are back to normal. I ran with Safari’s button tabs right before this update, when the change was foisted on my Mac in a Safari update. It was absolutely hideous. I’ve done everything I can to avoid using Safari, even setting Vivaldi as my default web browser. Between these results, the button-vs-tabs fiasco, and other issues, I now have no trust in Safari, nor the Apple developers who work on Safari, to produce a quality product.

So there you have it. A very brief look at four of the latest browsers on the current macOS. And before you ask, I will never install Microsoft’s Edge on my Mac.

running oracle virtualbox on macos 11.6 big sur

I run Linux virtual machines on my MacBook Pro. The MBP is a mid-2019 machine with an i9, 64 GiB of memory, and 4 TB of SSD. It is a powerful developer machine that has been in constant use since I purchased it last June.

One of the tasks I use the MBP for is running Linux virtual machines. For that purpose, there are two well-known tools; Oracle’s VirtualBox, which is open source and free, and Parallels Desktop, which is commercial. Both are capable of running most Linux distributions I care about, and Parallels is known for running Windows, and rather seamlessly, on the macOS desktop. The only problem is I have no need to run Windows as a VM, as the critical application Office is available as a native macOS application.

One major difference between VirtualBox and Parallels is that VirtualBox appears to need a kernel extension installed for its VMs to operate. VirtualBox ran just fine until I updated macOS to 11.6. It ran without issue on every release before that. When I stepped up to Big Sur 11.6, VirtualBox VMs would no longer boot.

Every time I would attempt to start a VirtualBox VM, I would get something like the following:

along with this:

Very annoying to say the least. I finally solved the problem thusly:

  • Boot the MBP into macOS recovery mode following these directions: https://support.apple.com/en-za/HT201314
  • When in recovery mode, open a terminal (Utilities > Terminal), using the menus on the upper left.
  • In the terminal execute csrutil enable --without kext
  • In the terminal execute spctl kext-consent disable

Then reboot out of recovery mode and back into macOS. The next time you attempt to start a VirtualBox VM, you will succeed.

I ran into another issue when I upgraded from VirtualBox 6.1.26 to 6.1.28. Under 6.1.28 I ran into the same VM boot issue, even though the documented changes were still in place. When I downgraded back to 6.1.26 it all started to work again.

If you’re wondering why I just don’t run everything under Parallels, it’s because Fedora and some of the other distributions I need to run don’t work very well with Parallels. The only distribution that works without issue under Parallels, regardless of version, is Ubuntu or Ubuntu derived distributions. Most of the time I’m quite OK with Ubuntu, but every now and then I need to I need to dip into Fedora.

My final thoughts on all of this: I spent six grand on this MBP. It’s my MBP, and as far as I’m concerned it will do what I want it to do, not what someone else feels it should be doing. Installing and running a kernel extension on macOS should not be a problem, and blocking it from installing is laziness on someone’s part who’ve decided perhaps it’s too much trouble to make sure that capability continues to work from release to release. I’m sure that trying to get a kernel extension running on top of Apple Silicon probably doesn’t work at all, which is why I’m not rushing out to get a new machine using it. I’m not interested in finding out that I can no longer dive into the internals and do what I want it to do.