Technology

A short trip through dependencies hell

With my recent embracing of Linux for my desktop PC, and my tendency to want to install various software that’s not necessarily ready for primetime, it was inevitable that I would run into the dreaded “dependencies hell” situation where something gets broken and ends up breaking other things. Well, I’m still typing this on my now-fixed Linux desktop, so this is definitely not one of these “screw Linux, I’m going back to Windows” post, and I’ll walk everyone through this incident in detail so that if by some chance you find yourself having a problem such as this, you will not panic and be able to repair things.

So yesterday I was going through some of my old archives and saw a big fonts package. We’re talking 1000+ fonts. I don’t want to install all of those, but I wanted to have a fonts manager app so I can have a visual preview of all of them and enable/disable some of them without wanting to move files around directly. One of the apps I saw listed as available for Linux was fontbase, and from the web site I thought it looked really interesting. It is distributed as an appImage, which I don’t like — we Linux users have to deal with multiple software packaging methods already — but it’s supposed to be a way software can be installed without requiring a separate app (like apt, flatpak, etc.) so I gave it a try.

Usually the way to install an appImage file is to make it executable and double-click it, so I do that… and nothing happens. I bring it up in Nautilus (“Files”), right-click and select run, and again, nothing. So I grumble and open a terminal from my Downloads folder and run it (with ./[file]), and get an error message:

dlopen(): error loading libfuse.so.2

AppImages require FUSE to run.
You might still be able to extract the contents of this AppImage
if you run it with the –appimage-extract option.

This error occurs because Ubuntu switched from FUSE2 to FUSE3. A long time ago. However I did not know that, and foolishly did not research this ahead of time. So I assumed that FUSE was not installed at all, and that’s how this all started.

Naturally I went to a terminal and entered “apt install fuse”. Apt warned me it would have to install a lot of packages and, more importantly, remove a lot of packages. However I was in the middle of playing a game and was not paying sufficient attention. If I had been paying attention I would have noticed that the packages to be removed included things like “ubuntu-desktop”, which is pretty important on a working Ubuntu desktop system!

Unfortunately I just hit enter to start the process. This did not have any impact on running applications so I kept doing what I was doing, playing the game (it was “Car Mechanic Simulator” on Steam, if you’re curious), and then I watched the Japanese Grand Prix, and went to bed.

When I woke up is when I realized how badly I had screwed this up. I woke my computer from sleep, but I couldn’t enter my password. So, I reluctantly did a hard reboot on the system — hold down the power key for 5+ seconds, then restart — and that’s when I became aware that I had made a huge mistake.

I made a huge mistake

So the computer restarts, I get the initial Ubuntu screen… and then it disappears and shows me the console output of the system startup. That’s not a good sign. Also the last message I saw was that gdm had failed. That’s really great, I thought sarcastically. And that’s when it hit me that installing the “fuse” package was the origin of the problem. I looked up whether the apt system has a rollback system, and it does not — maybe that’s why Ubuntu now uses snap. However it gave me a starting point. Fortunately I still had my phone working so I looked up fuse and realized that I had inadvertently rolled back FUSE3. So I switched to a terminal in text mode by pressing “Alt + F2”, logged in, and reinstalled FUSE3. However I also noticed that it didn’t involve installing a huge number of other packages but one thing at a time. I rebooted.

Of course I was still left in the system console after rebooting, with gdm loading but me not seeing a GUI. So again I switched to a terminal, and thought about what my next step was going to be, and fortunately I found that apt has a flag you can use to reinstall some parts of the system (–reinstall). So at that point I started looking for a “desktop” meta-package and immediately saw “ubuntu-desktop”, so I went for it with this command:

“sudo apt install –reinstall ubuntu-desktop”

and this operation took care of all the dependencies. I crossed my fingers, rebooted, and Hallelujah! I got my desktop back, and my first thought was “I gotta write this down”, that’s the technical writer in me I guess.

Fortunately Ubuntu has the ubuntu-desktop metapackage available. Had this happened with another distro I think recovery would not have been nearly as easy  And fortunately I remembered exactly what I had done at the system level which could have caused this problem. Because, make no mistake, I am the only person who caused that problem.

However it might be a good idea to “blacklist” the fuse package from getting installed by apt, as this really shouldn’t be done by anyone who’s running the Ubuntu desktop. I’m not exactly sure how this could be done, but it’d be a good idea.

That being said, and even better idea is to be careful AS ONE SHOULD BE when installing packages using sudo and apt, and to not just breeze through stuff because of a foolish sense of self-assurance that “I know what I’m doing”. And that’s ultimately the lesson from this — sudo as a tool is powerful, and the way it works on Linux is to add a step (authentication) that’s literally there to stop you from doing something stupid. In Breaking Bad the expression “respect the chemistry” keeps coming up. Well, if you’re using Linux, remember to “respect the permissions”. But, if you do get yourself into trouble, at least Linux is made in such a way that it’s possible to walk back your errors and get back to a working system, which is more than I can say for Windows.

A tool for file recovery cleanup

If you’ve ever run file recovery tools on a disk, you know that you can end up with multiple copies of recovered files. Well, I made a little script that can help reduce the number of duplicates for you to clean up.

https://github.com/eltopo1971/file-duplicate-nuker

fileDuplicateNuker takes a directory as an argument, then recursively goes through that directory and takes a hash signature from the files in it. When it encounters a file with the same hash signature, it deletes the file.

Does this take care of all the duplicates? Oh heavens no. That’s a feature, not a bug — call it erring on the side of safety. The script has no idea what kind of file it’s dealing with. All it does is take a hash signature and base the decision of whether to delete the file on that. If there is so much as one byte of difference in the file it’s examining, it’s counted as a unique file and not deleted.

That being said, from my testing it does delete a good number of files, and when you have thousands of files to wade through, any little bit helps.

Running a Windows virtual machine on Linux using an Existing Windows Installation

I’ve been a Linux user for years. At home I’ve kept a file server which runs off Linux, but that was “classic Linux” — without a graphical user interface, using the command line, and doing things the “hard way”. My main desktop PC was still running Windows, and had to because the one game I play regularly, Fortnite, was not available on Linux.

Well, someone recommended that I check out GeFORCE NOW, which is a virtualized environment you can use within Linux to play a large number of games in the cloud. I’ve always been skeptical about that type of play, largely because the capacity of the system to provide a good gaming experience had not previously been available. However since I live alone and have a good broadband connection I tried it out and loved it. It’s indistinguishable from playing the game from your own computer.

So, since most of the apps I use are open-source or otherwise available for Linux, I decided to switch from Windows to Ubuntu 24.04. My PC has two hard drives as well as a NVME drive which hosts the Windows system, so I decided to delete and transfer what I could from a 1TB drive and installed Ubuntu on there. It was a breeze and I was quickly up and running. Ubuntu has very good hardware support so most things just worked, and the only thing I had to hunt down was a driver for the Logitech G13 gamepad I constantly use. I’ve been running this for about 3 weeks now and I have no desire to go back.

However, today I came up against an obstacle. On Windows I used Bitvise SSH client to connect to this server, and Bitvise saves its files in its own binary format. I found myself in a situation where I would have to go back to Windows to use Bitvise to connect to my server. Also while my current situation does not require that I have access to a Windows machine, that can change all too easily. So I decided to create a Windows virtual machine (VM from now on), but instead of using a virtual hard drive file, I would simply use the disk on which Windows is already installed. That makes a lot more sense. It’s just more efficient.

I found one set of instructions to help me do that, but it dates back to 2021 and hasn’t been updated more recently, so I spend some time figuring out the more current way to go about it.

How to run a Windows VM from an existing Windows disk

For this you will naturally need a computer running Linux, preferably something Debian-based like Ubuntu. You will also need to install VirtualBox on that computer (downloads here). You should also disable Bitlocker encryption from your Windows drive before proceeding. I didn’t have it enabled on my drive so I don’t know how that would affect the installation.

Set up your user

Your user will need to be part of two groups: “disk” (to enable raw disk access) and “vboxusers”.

sudo usermod -aG disk,vboxusers [user]

Also let’s create a folder “vms” in your home directory in which you will keep your VMs:

mkdir vms

Set up your Windows disk

First, find out where your Windows drive is mounted. By default when you install Ubuntu on a computer with FAT32 or NTFS drives, they will be accessible to Linux.

lshw -short -class disk,volume

This will show you a list of your hard disks and their partitions. Look for a disk that contains a “Windows FAT volume”, a “reserved partition” and one or more “Windows NTFS volumes” and note the entry in the Device column for the disk (not partition). Usually this will be “/dev/nvme0” (if you have a NVME disk) or “/dev/sda”.

Next, we’re going to create a special file that points to that disk using a utility that is installed with VirtualBox. Enter this as your regular user (not root):

VBoxManage createmedium disk --filename=vms/[disk file].vmdk --variant=RawDisk --format=VMDK --property RawDrive=[Windows drive]

This file points at the location of the Windows drive.

Create a VM

Now we’ll start using VirtualBox itself. But before we do, let’s install the VirtualBox Extension Pack (download from here). To install the file just double-click on it on the file. It will launch VirtualBox and the installation will take place.

In the VirtualBox dashboard, click on the Create a new virtual machine (VM) link.

  1. In the VM Name field, enter a name for the virtual machine
  2. The VM Folder should be “/home/[user]/vms”
  3. In the OS field select Microsoft Windows
  4. In the OS Version field select the version of Windows installed on your Windows disk.Win12
  5. Click on the Next button at the bottom right of the New Virtual Machine window.
  6. Under Specify virtual hardware, adjust the Base Memory and Number of CPUs. Bring the Disk Size slider to the lowest value. Note that using 16GB (or more) of RAM is highly recommended otherwise you’ll find the VM experience very taxing, but keep in mind that this memory will not be available to your Linux apps while the VM is running.
  7. Select Use EFI
  8. Click on the Next button.
  9. Click on the Finish button.

Attach The Windows Disk to the VM

Now we’ll attach the pointer file we created in step 1 to the VM.

  1. In the Machines tab of VirtualBox Manager, right-click on the new VM and click on Settings.
  2. In the [VM]-Settings window, select the Storage tab.
  3. You’ll see a .vdi file which we won’t be using. Click on the Add Attachment button at the bottom right of the Devices box (see below) and select Hard disk from the dropdown menu. Attach Disk
  4. In the Hard Disk selector window, click the Add button.
  5. Select the .vmdk file you created earlier in the file selection dialog box.
  6. With the .vmdk file selected, click on the Choose button.
  7. In the [VM]-Settings window, select the .vdi file, and click on the Remove Attachment button (next to the Add Attachment button).
  8. Click OK to save your VM configuration.

Run Your VM

Nothing left to do but to run the VM and make sure it works, so in the Machines tab of VirtualBox Manager, right-click on the VM you just edited and select Start > Start with GUI.

You should be able to log into your Windows installation.

If you have tried in the past to install Linux and modified the UEFI partition of your boot disk… well, you will then have to navigate around the disk using the GRUB CLI to fix your boot sequence. This is beyond the scope of this particular tutorial, but instructions are easily found online. I had to do this myself.

Something else that came to mind as I was writing this was to try and see if I can do the same thing using QEMU instead of VirtualBox, which I will also write a tutorial for if I can manage to do it.

Keep in mind that virtualization at the local level can be a bit tricky and resource-intensive. It’s also one of the rare things that can completely freeze up your system and force you to reboot it — that’s called a kernel panic.

Elon Musk Keeps Spreading a Very Specific Kind of Racism

If nothing else positive comes out of his acquiring Twitter, it will at least have provided the public with an opportunity to see how obsessed Elon Musk is with race-related conspiracy theories.  A surprising amount of it is just plain creepy.

Elon Musk Keeps Spreading a Very Specific Kind of Racism (Mother Jones)

 

With the spotlight on him Elon Musk doesn’t look like much of a genius anymore

A highlight: “Musk is a right wing demagogue and a pathetic, narrow-minded racist, one that endorses causes that suppress the marginalized and champion white supremacy.”

The Descent of Elon Musk by Ed Zitron

What drives a man to “suicide”?

On March 9th a man named John Mitchell Barnett was found dead in a Charleston NC hotel parking lot, victim of an apparent suicide. But this wasn’t just any ordinary schmoe, Barnett was the main whistleblower for assembly and quality issues in the Charleston Boeing plant that produced the 737 MAX.

Funny how that happened. The man dedicated the last few years of his life to exposing problems that put the entire flying public at risk, and just when the issues he warned about are making headlines, suddenly, he “commits suicide”. Come on people. Sure, he died of a gunshot wound, but I would stake a large amount of money to say that he was not the one to pull the trigger.

The Charleston County coroner ruled the wound was self-inflicted, but when you think about the amount of pull that a huge employer like Boeing have on a place the size of Charleston you realize how the wheels of justice are sometimes greased just enough by major economic players into “being team players”.

Here is a video on the man and the major safety issues he tried to warn the public about.

YouTube player

It’s time to admit I overestimated humanity

Does the internet make people stupid? I mean, it seems like a silly idea. After all the internet gives you instant access to the entire knowledge of the world (along with a whole lot of BS) so surely that can’t be a bad thing… well, after seeing this screenshot I am not so sure.

A person on the internet asking if ramadan is a new tiktok challenge

Maybe AI isn’t ready to take over the world yet

Shared today on BlueSky — apparently this is the result of asking ChatGPT to illustrate what its core values are. So, there’s probably not much cause to be worried that ChatGPT is going to steal your job, because most jobs out there require communicating in, well, *a* language, and not some weird babble invented by randomly throwing syllables together.

ChatGPT values: weird icons with gibberish captions.

Twitter: a post-takeover poison pill

When a company is about to go through a hostile takeover, the stakeholders in the company have this strategy that’s available to them called a “poison pill”. The idea of the “poison pill” is that the shareholders, worried about the effects of the takeover on the long-term health of the company, will artificially depress the stock price of the company so as to make it unattractive for a takeover.

This is obvious *not* quite what’s taking place at Twitter right now.

@elonmusk‘s offer was so much over the realistic valuation of the company that the shareholders just saw $$$$ and went with it. However Twitter isn’t a traditional company. Twitter is a social network. Its technology stack is robust but it’s not particularly outstanding. It works, it doesn’t have a huge lot of features, but it can handle the traffic. Its real value is in the users and the connections it brings to the party. To remain at its baseline of “value” compared to before the takeover, it has to retain its userbase. If users leave, the site’s value is diminished. And this is something that @elonmusk
 doesn’t grok.

While he doesn’t get it, users do get it. And their response to Musk’s “comedy of errors” tenure ever since he took over the site is to look elsewhere for a new social network to spend time on, because it’s become clear that Musk wants to take this site and turn it into his personal sandbox. I wouldn’t pay $44 billion for a sandbox, but then I don’t have the sort of detachment from reality that being the world’s richest man engenders.

However for a couple of weeks now we’ve had a look at what @elonmusk considers entertainment for himself, and we’re all pretty much horrified, from Nazi imagery to petty personal fighting to non-stop lying by Musk himself. And that’s why, sadly, now is the time to ditch this platform. Because remaining a part of it at this point is to risk immeasurable personal reputational damage. Think the repercussions in your life if it came out that you were a user of “stormfront” (or whatever KKK-affiliated web site exists out there). This is what Twitter will turn into in the hands of a spoilt man-child with highly questionable morals and a reputation as a con man who has no board to answer to and in time is growing more and more embittered that he can’t just buy a positive image for himself. Or friends.

And if that sounds like I’m describing Donald Trump, it’s not a coincidence; both Trump and Musk are trust fund babies whose lives are led by their malignant narcissism. 

So there’s an understandable urge to leave a platform that’s devolving into a giant cesspit of xenophobia in all its diseased forms, because users don’t want the taint of it.

It’s probably a bad idea to deactivate one’s account, however. All this will do is leave your handle open to a malignant actor taking it over and attempting impersonation. A much better approach is this: make sure you set multi-factor authentication on your account, and then log off. This way no one can use your handle, and you are protecting your reputation.

There are many alternate social networks out there that don’t belong to snake-oil-selling egomaniac billionaires, such as mastodon, counter.social and tribel. Check them out and give them your time and eyeballs instead of watching someone who should know better tank a platform to flatter his own malignant ego.

What is a blockchain?

2018 is poised to be year when cryptocurrencies become mainstream. The original cryptocurrency, Bitcoin, has entered the common jargon of the modern world last year as its valuation hit record a record high of nearly 20k USD/BTC, and stayed in the news as its valuation dropped to more reasonable levels. Ethereum is also gaining recognition as it became the #2 cryptocurrency in terms of market capitalization. In short, a little over 8 years since the creation of Bitcoin cryptocurrencies are gaining recognition and acceptance in the “real” world.

Cryptocurrencies are created as part of something called a blockchain. And more than cryptocurrencies, it is the blockchain idea which is expected to have a huge impact on the computing world, at least for the next couple of years. As such it is a good idea to learn what a blockchain is, at both a basic and more advanced level.

The Basics

At its core, a blockchain is a distributed ledger. Those with an accounting background will immediately recognize what a ledger is — it is a record of transactions. A blockchain is distributed, which means that entries in the ledger are written by many parties, as opposed to by one centralized authority.

Like an ordinary paper ledger, blockchains are write-once. Once a block has been verified and added to the blockchain it cannot be erased or modified. This insures that transactions cannot be taken back.

The Nodes

All these “parties” are actually computers running a node for the blockchain’s network on the internet. This involves executing software which contributes to the blockchain network. Depending on the network involved there may be several types of nodes in a blockchain; this will be explored in depth later.

The Blocks

Nodes compile a number of transactions into a block. How large the blocks are, and how often they are verified, varies widely between blockchains. For example, the Bitcoin blockchain generates a block every 10 minutes. The Ethereum blockchain, in comparison, generates a block in less than 20 seconds, and Bitshares blocks are generated every 3 seconds at most. A number of factors affect block time; if you’re not intimidated by math check out this article for more information.

The Chain

Blockchains are so named because each new block is appended to the previous block, effectively forming a chain. In fact one can always look at certain information in the latest block of any given blockchain and trace the blockchain’s history all the way back to its very first block.

Hashing

Since blocks are appended to the blockchain by several different nodes, there needs to be a way to ensure that only the block with the right data can be added at any given time. Otherwise there would be no way of ensuring the continuity of the blockchain from the genesis block to the most current one.

This is where hashing comes in. Hashing is a cryptographical technique that is used to generate a unique code that can be used to identify a set of data, rather like a fingerprint. The hash is generated from the transactions contained in the block and recorded as data in the block, which also includes the hash from the previous block. This is one of the mechanisms used to verify any new blocks. If the previous-block hash does not match the previous block’s recorded hash, then the current block is invalid and cannot be added to the chain.

The actual library used to generate the hashes depends on the blockchain. SHA256 is a popular one and is used by Bitcoin. Other libraries include scrypt, X11, Cryptonight and ETHash.

Hashing produces a completely different string if there is any change whatsoever to the original hashed content. The SHA256 library can produce a very large number of distinct values (3.4028237e+38) so arriving at the same value from two different pieces of content is extremely unlikely. By comparison, the chances of winning the Powerball lottery in the USA is 1 in 2.92e8. One could win this lottery 4 times and that would still be less likely than generating the same hash from 2 different sources. Thus the use of hash values makes blockchains virtually tamper-proof.

This was a very basic overview of blockchains. We’ve barely scratched the surface. In my next few articles I will be providing more in-depth coverage on subjects such as concensus algorithms, blockchain node types, the relationship between blockchains and cryptocurrencies, and how the blockchain can be used by businesses to streamline processes and reduce processing costs.