Adam Labay

Feb

2020

Taking Screenshots on the PinePhone… sort of

We all know that, when playing around with a new device, the most important thing you can do is record the experience so that others can behold your techno-wizard glory. Pics or it didn’t happen, no? Which makes it that much more frustrating that the PinePhone currently can’t take screenshots in seemingly any of the OSes that are compiled for it.

Attempt 1: The Built-In Option

I’ve been using Ubuntu Touch, since at the moment it has the most features and the least suck of the available OSes. (Technically the build is from UBPorts, but I’m going to use the terms interchangeably because I’m lazy.) Allegedly, UBPorts lets you take a screenshot by simultaneously pressing Volume Up and Volume Down, but that doesn’t work on the PinePhone – best I can gather, that only works on Android ports of Ubuntu Touch.

Attempt 2: The Command Line Option

Other posts suggest that you can use phablet-screenshot from the phablet tools package, but that too leads to disappointment, with the program freezing. Running debug mode (-d) shows that it’s trying to pull the screenshot from adb:

...
+ adb start-server
+ check_devices
+ set +e
+ adb wait-for-device

Since the PinePhone isn’t running Android, adb isn’t really a thing, and you’ll end up waiting forever.

Attempt 3: The Raw Option

Looks like we’re going to have to create this from scratch. All of the above-mentioned techniques essentially do the same thing: they dump the framebuffer /dev/fb0 to a file and then convert that file to PNG, JPEG, or whatnot. How hard can that possibly be?

Dump the Framebuffer

The first step is easy enough: cat /dev/fb0 > img.raw. This dumps the framebuffer (i.e. what’s on screen) to a raw file.

Figure Out Your Screen Size

The raw image file is just that: a raw stream of pixels, with no dimensions built in. As such, you next need to figure out how big your screen is. For the PinePhone, this is 720×1440 pixels. If you want to discover this yourself, you can just type fbset (sudo apt install fbset if necessary).

Get the Script

This helpful chap cooked up a lovely little Perl script that will then convert the raw image to png. Download the script, chmod + xto mark as executable, and invoke with ./iraw2png 1440 720 img.raw img.png.

NB: I swapped the dimensions. You’ll see why.

Inspect the Results

Here’s the result, which doesn’t exactly look like the Desktop:

Success!

It definitely captured a thing.

Mope!

It captured some manner of console, perhaps the TTY output? I don’t know. Evidently there are other framebuffers to be found. If you know how to find them, please do send a comment since I’m up a creek on this one.

Feb

2020

Mistakes to Avoid With the PinePhone

As mentioned before by me and many, many others, the PinePhone Brave Heart edition is very much not road ready. That’s fine, but it does mean you’ll probably find yourself reflashing your device’s image a few times. What follows are the various mistakes that have led me to re-flash my phone.

Decreasing the Brightness

As soon as you move that slider, your brightness will go to nil. As in no image at all. If you have SSH enabled, you can probably set that brightness back, but there’s a good chance that you’ll make the move prior to having SSH set up. Whoopsie.

Applies To

Sailfish OS
Ubuntu Touch (only if you decrease it fully)
PostMarketOS (only if you decrease it fully)

Turning off WiFi

~~Under Ubuntu Touch, turning off WiFi somehow borks the radio, requiring you to toggle it off and back on. Not a huge nuisance, but not nothing either.~~

Applies To

~~Ubuntu Touch~~

Following that modem.txt document

A lot of folk, myself included, have tried to enable voice calls by following this file.

Don’t. do. it.

The file relies on a custom kernel, and if you set your QDAI settings per this file, you will end up unable to make phone calls until you reset it to factory defaults – which you have to do manually.

Feb

2020

Ubuntu – The Most Stable PinePhone OS So Far

Of the four operating systems I’ve tried so far on the PinePhone, Ubuntu Touch (aka UBPorts) is the most stable, getting me nearest to actually placing a successful call.

Installation

The process starts by downloading the latest image, either from the Pine64 wiki or from the UBPorts Jenkins site. Flash the image onto a microSD card using Balena Etcher or your favorite flashing tool.

First Use

Initial startup works like a charm. You’ll want to attach to WiFi since the SIM won’t initially be read. Once you’re in, hop on the Interwebs and have some fun. Just nothing that requires audio, since that’s the next step.

Enable Mobile Data

Ironically, your fancy new phone won’t be able to use any mobile functions. You need to enable the modem, this time and every time you boot. There’s probably a way to turn the process into a startup script; I’ll report back if so.

At any rate, open the Terminal app and enter these commands (taken from this post):

sudo /usr/share/ofono/scripts/enable-modem
sudo /usr/share/ofono/scripts/online-modem

The sudo password is phablet

Once you’ve entered the first command, the second one is a piece of cake thanks to Ubuntu Touch’s nifty keyboard. Swiping up on the keyboard reveals the cursor, which is moved by swiping around. Nice touch, UBPorts.

Enable Audio

Much like enabling mobile data, you’ll have to do this every time you start the device.

Open the Terminal app and enter the following:

sudo modprobe snd_soc_simple_amplifier
sudo modprobe snd_soc_simple_card_utils
amixer -c 0 set 'AIF1 Slot 0 Digital DAC' unmute
amixer -c 0 set 'Line Out' unmute

Optionally, add the line amixer sset 'Line Out' 100% to max out the volume (best I can tell, you’re setting the volume that 100% on the volume rocker corresponds to; i.e. if you leave this at 55% (the default), the loudest you can make your phone is 55% of its true potential).

Don’t Disable Wifi

If you do, it won’t come back. Rather, it will appear enabled, but no access points will appear and even saved APs won’t connect. To remedy, open Terminal and enter the following:

sudo nmcli radio wifi off
sudo nmcli radio wifi on

What Works

So far, I have successfully

Played YouTube videos
Surfed the web on mobile data
Rebooted the phone a bunch of times

What May Yet Work

It turns out that my cheapo FreedomPop SIM requires some proprietary app to do text and calls, so I can’t test that at present. But if the Reddits and the forums are any indication, I’ll be a lucky son of a gun if I manage to successfully place a call. Texts shouldn’t be a problem.

What Doesn’t Work

Taking a friggin screenshot. Maybe there’s another app required?

Updates to follow.

Feb

2020

PinePhone First Impressions

My PinePhone BraveHeart edition arrived yesterday, and the past few hours have been spent playing around with it and trying to place a basic call. Findings so far:

For Developers and Maniacs Only

The folks at Pine64 make very clear that this phone is not a daily driver, but it bears repeating: this is not so much a phone as a fun Linux project with a SIM slot. You have to install your own operating system, and none of the choices on offer claim to be even remotely road-ready. In particular, the four I’ve installed so far have had one or more of these symptoms:

Unusably slow
Can’t find the SIM
Require a light to be shined on them to display anything
Lock you out without a password

Further, even the folks online who have managed to get past the above nuisances still generally can’t get the silly thing to successfully place calls – you know, the one thing a phone is supposed to do. So again, not a daily driver.

But All Kinds of Fun to Play With

With all those caveats out of the way, the thing’s a hoot to mess with. If something goes pear-shaped, just pull out the microSD card and re-flash it. And at $150, the price point is great for a project.

What Lies Ahead

I’ll do my best to make the silly thing work, and will report back with whatever tips and tricks I come across.

Feb

2020

Adventures with XYO – The Thrilling Conclusion

Back in November I decided to try out mining XYO using the Coin app on an old Nexus 5X that I reappropriated for the purpose.

Now it's February, and I think I'm about done with the whole thing.

To be candid, the main reason is that my sentinel went through the wash, and while it looked like a fresh battery would fix it back up, the new batteries don't seem to last more than a couple weeks. As such, I'm out $12.95 for the sentinel and $7 for a set of replacement batteries. In the same amount of time, I've earned about 8,000 COIN, which, if I could convert them to XYO (that option doesn't start until 10,000), would be worth...

31 cents.

I'll probably throw a few other posts about the various aspects of the COIN app and geomining, but for now here are my main takeaways:

Great for Truckers

Perusing the Reddits, the one group of people who seem to get a decent-ish return from geomining are truckers, who spend literally all their time cruising from tile to tile. I'd wager that Uber drivers fit this description as well, though truckers have the advantage that they can geoclaim gobs of regions every cycle. Truckers, who pass from cell to cell quickly and thus need a rapid recharge cycle, are probably the only group who benefit from Pro accounts as well.

And Nobody Else

As an ordinary schlub who works in an office and either drives to work or takes a train, there's simply no return to be had. Five miles of walking net about 100 COIN, which isn't a whole lot greater than just sitting around refreshing the tile periodically while I work from my desk. Geoclaiming isn't an option either, since the two regions I can claim (work and home) are both chronically oversubscribed, making the a losing proposition for everyone involved.

It's a shame, really. The premise of geomining, as outlined on an XYO Medium post, ain't bad. Using a fleet of smartphone-armed passers-by to determine whether packages arrived is clever, and I really do hope they figure out a way to make it work. For now, though, it's just another GPS device attached to yet another freight truck.

Nov

2019

The Futility of Geoclaiming

As described by the folks at the COIN App:

Geoclaiming

When you Geoclaim a tile, you are becoming an owner of the tile for the remainder of the campaign(Currently UTC Monday to UTC Sunday). Anyone that Geomines using the COIN app in your area will give you a chance, based on your Ownership percentage, to collect 10% of rewards.

An Owner with 100% ownership of a Geoclaimed Area will receive the 10% of rewards every time.

An Owner with 50% ownership of a Geoclaimed Area will receive the 10% of rewards half of the time.

An Owner with 25% ownership will receive the 10% of a miners reward a quarter of the time.

In theory, this is a way of earning that sweet, sweet COIN without having to actually, you know, go places.

In practice, it’s a lovely exercise in overestimation and accidentally returning COIN to the system.

But it doesn’t have to be, and my current goal is to figure out means of optimizing geoclaims.

The Basics

The amount you get out – your return – is proportional to your stake – i.e. the amount of the area that you control, or what the COIN folks call your ownership.

I’m using stake because in this context, ownership is a misleading term. Take a look at this screenshot:

Throughout this week, my ownership has never dropped below 10%, and has usually hovered in the 13-14% range. Yet, my returns are well less than even 10% of the total 10,940 COIN held by this region.

That’s because you don’t own a fraction of the COIN held; you get a fraction of the COIN earned, as it gets mined. And remember that geomines are random numbers, further dependent on whether that user happens to have a sentinel at the time. It seems that my RNG luck this week was pretty crap, since the users I was earning from tended to have lower mines than average. That, or a lot of users in my area have premium accounts, which prevents their mines from entering the pool. Probably both.

The Point of No Return(s)

Say it with me: your earning potential is not unlimited. It’s determined by your stake, and by the available earnings. If we make the simplifying assumption that your stake is actually an ownership, then

the amount you pay into the geoclaim – your investment – must be less than your stake multiplied by the total COIN in the region.

Seems obvious, but you would be amazed how many people on Reddit get hyped about how much they dumped into their latest geoclaim without realizing this basic fact. And, as discussed earlier, your expected return is actually significantly less. (How much less? I’m working on a heuristic, but that’s another post.)

Consider the screenshot above: That 13% stake cost me 1905 COIN. Even if I earned a full 13% of the available ~11k COIN, I’m still not going to recoup my investment. (In this case, I saw that coming, but dumped the all the COIN in for other reasons.)

Good Money After Bad

But what if I just increased my stake until the math worked out? There’s a possibility that would work, but it depends on two things:

First, the distribution of investments. If it were linear, then doubling my investment would double my return. But it’s not linear, it’s a Pareto distribution (more on that in a future post). TL;DR: it’s like the American economy, where most of the stakes are taken up by a very small fraction of users.

It turns out that doubling your investment does indeed double your return. The only thing that matters is how much has been invested total, below.

Second, there is a point of no return. If the total investments ever exceed the available COINage, then nobody gets a positive return, no matter how much they put in.

The math is easy enough: Again, pretending that your stake is an ownership – which, emphatically, it ain’t – we can say that your return is defined by

$R = s \times A$

where R is your return, s is your stake, and A is the total available COIN. Your stake, meanwhile, is defined by the ratio of your investment to the total investment:

$s=\frac{i}{I}$

So, we end up with

$R=\frac{i}{I} \times A$

If we want a positive return, then

$\frac{R}{i} > 1$

meaning

$\frac{A}{I}>1$

or, the total available COIN must exceed the total investment. If everybody tries to pump up their stake, then there comes a point where everybody loses. Whoops.

So can Geoclaiming work?

It depends. Specifically, it depends on the distribution of stakes, and a few other factors. I’ll get back to you.

Nov

2019

Adventures with XYO

For the unfamiliar, XYO is yet another cryptocurrency venture. The basic premise is that users mine their coin (also called XYO) using the Coin app, which uses proof of location as the mining function (as opposed to proof of work for Bitcoin, or say proof of storage for Burst or Filecoin).

You wander about the world, which is divided into squares about 25 feet on a side. Each square you enter is worth a random amount between 0.02 and 0.1 XYO, which gets “geomined” when your phone enters that region. If you pay them $12.99, you can get a Sentinel – a bluetooth doodad that, as a reward for foolishly dropping 13 bucks on their technology, multiplies each geomine by a factor of 12.

This method of mining, however, is crap. As an average person who walks around casually (about 6 miles a day, much of it around the office), a daily haul is ~200 Coin (not 200 XYO; read on.). At that rate, it would take about 2 months before I could convert the minimum 10,000 Coin into 10,000 XYO, which as of this writing is worth about $2.95.

As in any good cock-and-bull crypto scheme, there are other ways to increase your rewards. One is a Premium subscription, starting at $24.99 a month, which multiplies all of your mining by various factors. Another is to hijack the mining of others by staking a Geoclaim – i.e. paying into a 25×25 mile region in exchange for a percentage of the coin mined therein. This comes with its own risks and rewards.

For funsies, I’ve decided to spend the next two months answering the question, Is it possible to actually generate any manner of money using XYO? I’ll wander about, rapidly draining the battery on the old Nexus 5X I’ve dedicated to the endeavor, do the math, and report back.

Nov

2019

I got hacked

The title says it all. No idea how, since I use 2FA on Dreamhost, but here we are. Time now to rebuild.

Aug

2019

Raspberry Pi 4, Kodi, and Chrome – An Uncomfortable Alliance

For years, I have needed two devices in order to watch quality drinking-time programming in my bedroom:

A Chromebit, which runs Chromium, for those late-night binges of TVTropes.org, but doesn’t run Kodi
A Nexus Player, which runs Kodi, among other things but doesn’t run Chromium
Two remotes and an HDMI switcher to make it work

Each component is fantastic at what it does – the Nexus Player, despite being largely mothballed, has a great interface, and the Chromebit is the easiest way to get Chrome on HDMI – but it sucks for the need to use two devices.

This summer, I decided to fix that mess.

What follows is a short list of things I tried, none of which could pull it off:

Raspberry Pi 3 running Kodi (too underpowered)
Raspberry Pi 3 running Emteria OS (see above)
Essential Phone running Android TV Launcher (mobile browser on a big screen is a wreck)
Raspberry Pi 4 running Emteria OS (not until Emteria releases a build for the Pi 4; I couldn’t properly hijack the Pi 3 build

Here’s what worked:

Here’s how to do it.

Step 1: Download and Install Raspbian

Download Raspbian Buster. I wanted to customize my GUI, so I went with the Lite ISO, but ain’t nothing wrong with the Desktop version.
Flash the image to your MicroSD card using your favorite flashing software. (I’m on Manjaro, so this meant BalenaEtcher, but Rufus is great for Windows.)
If you’re on WiFi, you’ll need to connect. This is trivial if you downloaded Raspbian Desktop, since it happens on boot. If you’re using Raspbian lite, enter sudo raspi-config and choose Networking Options. Enter the SSID and password.
Enable SSH. If you’re using Raspbian Desktop, then open a terminal window (if you’re using Lite, you’re already at one). Enter sudo raspi-config, choose Interface Options | SSH and enable.
Reboot to make sure the change stuck. You can tell if the WiFi setup worked by typing ip address to confirm that you have a proper IP, and you can confirm that SSH is installed by ssh’ing into said IP.

Step 1b: Install a Desktop Environment

If you used Raspbian Lite, you’ll need to install a desktop environment. Everybody has their favorite and I’m not going to take that away from you. I went with XFCE because it has rather beefy window icons and ready support for changes to DPI. Both will be important since I’m watching this on a projector and navigating from my bed, so usability is paramount.

To install, just type sudo apt install xfce4 and let the magic – and about a million dependencies – happen.

I would also recommend sudo apt install network-manager-gnome. This could be unnecessary, but I found it stopped having the WiFi stop working after the first or second reboot. Admonish in the comments if this is unnecessary.

Step 2: Install Kodi and Chromium

Straightforward enough:
sudo apt install kodi chromium-browser seahorse

Kodi has the ability of starting as its own desktop environment. We will eschew this option, since the only advantage of doing so is that you don’t see the Applications bar when you hover over it. Meanwhile, the disadvantage is that you can’t run any other Gnome apps, which is irritating.

Step 3: Log into XFCE, make it the default, and suppress them password dialogs

Reboot (sudo reboot), and when logging in, make sure that XFCE4 is selected as your desktop environment (top-right corner). Logging in is annoying, so we will invoke auto-login.

Similarly, when opening Chromium, entering a password for the default keyring is a pain, so we will suppress it.

Both of these moves are strongly advised against by StackExchange pedants, and that’s fair enough. But this is a media center, not the NSA, and convenience is paramount. Further, we’re going to change the pi’s password so there’s at least one layer of security still present.

Open a terminal window by clicking Applications | System | XTerm. Type passwd and follow the instructions to change the default password from raspberry to something else.
Run sudo raspi-config to configure auto-login: Select Boot Options | Desktop / CLI | Desktop Autologin and then Finish.
Exit the terminal. Open Seahorse password manager by clicking Applications | Accessories | Passwords and Keys
Right-click on Login and choose Change Password. Change the password to <blank>. This will disable prompts to unlock the default keyring.

Step 3b: Optionally, Configure HiDPI Mode

Tiny buttons are hard to hit with a remote control, and tiny text is hard to read when projected across the room. The easiest way to uniformly scale everything on screen is to adjust the DPI settings. This is optional and depends on your use case.

Open Applications | Settings | Appearance and under Fonts change the DPI to 120 – or whatever works for you.
Open Applications | Settings | Window Manager and under Style select Default-hdpi
Reboot to see the full effects.

Step 4: Configure Advanced Launcher

Advanced Launcher is a Kodi addon that creates icons for ordinary apps. We’ll use this, plus a shell script, to launch Chromium from within Kodi

SSH into your pi
Download Advanced Launcher by typing
wget https://github.com/SpiralCut/plugin.program.advanced.launcher/archive/master.zip
Type nano chromium.sh and enter the following file contents:
#! /bin/bash chromium-browser --user-agent="Mozilla/5.0 (X11; CrOS armv7l 10895.56.0) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/69.0.3497.102 Safari/537.36" --window-size=1920,1080 $1
Return to the pi and open Kodi. Install Advanced Launcher.
Navigate to Add-ons | Add-On Browser (the open box at the top next to the Settings gear)
Choose Install from Zip file and navigate to Home Folder and select master.zip. You’ll likely have to answer a prompt to enable unknown sources.
Advanced Launcher should now be under Program Addons in your Addons folder. Open it and navigate to Default. A dialog to create a new launcher should appear.
1. Choose Standalone Launcher. Navigate to Home Folder and choose chromium.sh
2. Leave arguments blank
3. Keep the title as Chromium – or change it. Free country.
4. Select Linux as the platform
5. Skip Thumbnails and Fanart
You now have a launcher that invokes Chromium from within Kodi. To make it more accessible, right-click on the launcher and choose Add to Favourites.

Step 5: Have a lovely beverage

There are additional steps if you want to use Netflix, Amazon Prime, or (heaven help you) Xfinity. But, at this point, you have a nice core setup, so lean back with a lovely beverage and celebrate your accomplishments. The rest can wait for later posts.

Jul

2019

Ubuntu Cosmic on Pixelbook – Possible!

Step 1: Create a chroot with Xenial

Ubuntu 16.04 Xenial, despite being obsolesced years ago, is the most recent version of Ubuntu that can be installed directly with crouton. To get to Cosmic, we’ll have to upgrade a few times. This first section is basically an excerpt from the official instructions at https://github.com/dnschneid/crouton

If you haven’t already, download crouton from https://goo.gl/fd3zc
Press ctrl-alt-t to open a crosh shell
Make the installer by typing sudo install -Dt /usr/local/bin -m 755 ~/Downloads/crouton
Install by typing sudo crouton -t xfce -r xenial -n cosmic
Follow the prompts

Congratulations! You now have Xenial installed. Enter by typing sudo startxfce4 -n cosmic

Step 2: Upgrade to Bionic

Now that you have a functioning chroot, it’s time to upgrade to Bionic – the most recent version of Ubuntu which is still fully functional on a Chromebook.

Enter your choot by typing sudo startxfce4 -n cosmic
Enter a terminal by typing ctrl-alt-t
Install Release Upgrader by typing sudo apt install ubuntu-release-upgrader-core
Initiate the upgrade with the following commands:
sudo apt update sudo apt upgrade sudo apt dist-upgrade sudo do-release-upgrade
Follow the various prompts
Log out, restart your Chromebook, and enter your chroot to make sure everything worked.

If all went well, you have yourself a fancy new Bionic installation.

Step 3: Replace a bunch of drivers

If you try to reproduce the above steps to install Cosmic, you will discover that you can never get a clean apt upgrade. Instead, there’s a set of 5 packages that keep being held back:

The following packages have been kept back: xserver-xorg-core xserver-xorg-input-evdev xserver-xorg-input-synaptics xserver-xorg-input-wacom xserver-xorg-video-intel 0 upgraded, 0 newly installed, 0 to remove and 5 not upgraded.

If you try to get rid of them, you’ll end up cutting off your display, which is sadness. Instead, we need to replace them with different versions, and then revert to these versions post-upgrade.

Remove the troublesome drivers with
sudo apt remove xserver-xorg-core xserver-xorg-input-evdev xserver-xorg-input-synaptics xserver-xorg-input-wacom xserver-xorg-video-intel
Install the friendly drivers with
sudo apt install xserver-xorg-core-hwe-18.04 xserver-xorg-input-evdev-hwe-18.04 xserver-xorg-input-wacom-hwe-18.04 xserver-xorg-video-intel-hwe-18.04
Type sudo nano /etc/update-manager/release-upgrades and change prompt=lts to prompt=normal
Execute the upgrade with
sudo apt update sudo apt autoremove sudo apt dist-upgrade sudo do-release-upgrade
Do not reboot after the upgrade! You will lose your video.
Reverse steps 1-2:
sudo apt remove xserver-xorg-core-hwe-18.04 xserver-xorg-input-evdev-hwe-18.04 xserver-xorg-input-wacom-hwe-18.04 xserver-xorg-video-intel-hwe-18.04 sudo apt install xserver-xorg-core xserver-xorg-input-evdev xserver-xorg-input-synaptics xserver-xorg-input-wacom xserver-xorg-video-intel
Put back the packages that were autoremove’d earlier:
sudo apt install libcurl3 libegl1-mesa libglu1-mesa libgtop2-common libnih-dbus1 x11-apps x11-session-utils xfonts-base xfonts-encodings xfonts-scalable xfonts-utils xinit xinput xserver-common

Now you can exit, reboot, and log back in.

Congratulations! Also, your system is now unusably slow.

Good news: you upgraded to Cosmic. Good job, give yourself a hug.

Bad news: It’s gonna be slow. Mad slow. Unusably slow.

But it was a fun academic exercise, wasn’t it? Remember, we don’t upgrade software to use it, we upgrade because it’s there.

Attempt 1: The Built-In Option

Attempt 2: The Command Line Option

Attempt 3: The Raw Option

Dump the Framebuffer

Figure Out Your Screen Size

Get the Script

Inspect the Results

Success!

Mope!

Decreasing the Brightness

Applies To

Turning off WiFi

Applies To

Following that modem.txt document

Installation

First Use

Enable Mobile Data

Enable Audio

Don’t Disable Wifi

What Works

What May Yet Work

What Doesn’t Work

For Developers and Maniacs Only

But All Kinds of Fun to Play With

What Lies Ahead

Great for Truckers

And Nobody Else

Geoclaiming

The Basics

The Point of No Return(s)

Good Money After Bad

So can Geoclaiming work?

This summer, I decided to fix that mess.

Here’s what worked:

Here’s how to do it.

Step 1: Download and Install Raspbian

Step 1b: Install a Desktop Environment

Step 2: Install Kodi and Chromium

Step 3: Log into XFCE, make it the default, and suppress them password dialogs

Step 3b: Optionally, Configure HiDPI Mode

Step 4: Configure Advanced Launcher

Step 5: Have a lovely beverage

Step 1: Create a chroot with Xenial

Step 2: Upgrade to Bionic

Step 3: Replace a bunch of drivers

Congratulations! Also, your system is now unusably slow.

Recent Posts

Recent Comments

Archives

Categories

Meta

Other Stuff I Produce