Tag: geek

Upgrading to Ubuntu 24.04

Didn’t go smoothly. I was on 22.04 I believe and I decided to upgrade without waiting for 24.04.1. It was a mistake. I started the upgrade with do-release-upgrade -d, but way too soon I ran into a very nasty crash screen telling me “Oh no! Something has gone wrong and the system can’t recover”.

Aside. I personally hate these “friendly” additions like “Oh no!” Fuck off? My system just crashed, I’m pissed and the last thing I need now is some linux dev who thinks they know how to be friendly rubbing it in.

The way to fix it for me was: first reboot and choose recovery mode in the Grub menu. Once recovery mode is on, enable network then drop into a root terminal. Run dpkg --configure -a to fix whatever was interrupted. After reboot it should boot all right (at least the GUI would work).

However the network was still broken. I could ping so the connection itself was alive, but nothing could be resolved. Turns out the installation removed the resolve component of systemd. Luckily for me sudo apt-get install systemd-resolved could install it back and that solved the connection issue.

At this point the system seems to be functioning properly (nothing is obviously broken other than Thunderbird for some reason getting removed from my dash), but I don’t know if the interrupted install broke anything nor how to check if anything is broken or in an uexpected state.

Bye Seafile

Operating Seafile was a very frustrating experience. In the first place, setting it up on my Kubernetes (k3s) cluster was quite a ride. It didn’t have a Helm chart or any manifests ready to spin it up, so I had to build my own. This turned out to be pretty difficult because it was clear that the software was not designed with containers in mind and it took significant amount of hacking to even just get it to work. But once it did, then you’d need to manually run garbage collection every few months, because that was a Paid Feature™. I put up with it because of the sunk cost fallacy and because other than that five minutes every 3 months or so, it was working okay. Until it wasn’t.


I’ve been trying to use linux for pretty much as long as I’ve had a computer. I’d kept running into blocks though: at first I was using my computer mostly for gaming, and that’s still not a strong suite for linux 20 years later. Then my workflows were dependent on tools that only worked on Windows (like Office’s Publisher, or Dreamweaver and Fireworks—hell I still miss the productivity I had with Fireworks). Then I had a laptop half of whose hardware wasn’t supported by any linux distro at that point. Then gradually those issues went away and I’ve been an Ubuntu main for over ten years now. Some linux elitists will look at my desktop and hiss that it’s not just a terminal or a tiled window manager that looks like it’s still 1995.

Configuring MPV

I used VLC for a very long time. No longer. The other day I was watching something and I noticed how dull the colors were. It did fit the theme of the video though so I just assumed that it was a thing, except then I ran into a very colorful screenshot from the same video. What gives? In 2011, when fansubbing was huge, 10bit was the “new big thing.” It encodes colors into 10 bits into the standard 8, which results in much better colors. At least it should, as it seems VLC wasn’t able to handle that.

Example of 10bit difference

A way to extract an image from Firefox

The problem I faced: I was scrolling through a hashtag‘s timeline in Tweetdeck, when I spotted a nice image. I wanted to retweet it, but that failed: turns out the tweet was deleted in the meanwhile. This usually means that either the artist themselves deleted the tweet for some reason, or it was stolen in the first place and (by some miracle) it was moderated.

In this case it seems to have been the latter, so finding the image through the author’s profile was not an option. As Tweetdeck displays images as CSS backgrounds in the timeline, right-click “Open image in new tab” was not an option. Since the tweet was deleted, its URL was invalidated too, so I couldn’t give that to Saucenao or similar to find matches either.

However, while digging through the Firefox Dev Tools, I noticed that the context menu when right clicking the URL of the background-image had an item “Copy image data URL”. This turned out to be just what I needed.

Pasting the copied (data blob) URL to the address bar I got the image that I could magnify so that I could decipher the artist’s signature on it. (It turned out to have been a partial image so Saucenao couldn’t help either.) After a few tries I found the artist’s profile and there the original post. Yay.






スマート電球っていいね… と思っちゃった。わざわざベッドから起きなくても消せるし、スマホから操作もできるし、遠距離で点けたり消したりできるし、明度も調整できるし、強いでしょう?とか思いながらも結局金出すほどやりたいことでもなかった。そして今年の夏、ようつべでルームツアーやインテリア動画にハマった時期に紹介の動画を見たら結局衝動買いしちゃった。なぜかというとその辺飲みに行った店の内観がめっちゃササって、それっぽくできるじゃない?と思ったから。

Dual screen workspaces in Ubuntu 17.10

I expected the aardvark to break a couple of things. They switched back from Unity to Gnome, so even more than usual, I feared the system might not even start.

Compared to that, things went quite smoothly. One problem I ran into is how Gnome devs apparently never used a dual screen setup, and thought it would be a good idea to only switch workspaces on the main screen, and keep the other screen static. Needless to say, that’s not how workspaces are supposed to work.

Android and Unicode ID3 tags

For a while now I’ve been noticing that certain songs’ ID3 tags are broken when played on my Android phone. I use Black Player which in turn uses the built in Android music libraries – but I’ve checked in a number of other players too just in case and the problem persists.

My music library is extremely multilingual (and most of those languages I don’t even speak), so it’s full of UTF characters. Some of them seem to break Android’s encoding recognition. Sadly some of these triggers are pretty common, resulting in borked last.fm scrobbles. (And of course last.fm can’t be expected to be so smart as to fix all those automatically.)

Adding an SSL domain

Setting up SSL at first was a laugh. The other day though I fell victim to a sale where they gave away 10 year .tech domain registrations for a measly $40. Considering that’s basically 90%, I enthusiastically picked up a personal domain for future use.

Then again I don’t have plans of setting up a personal portfolio site for now (though I used to have one a decade ago), so I wanted to just redirect it to valerauko.net instead. That’s where things got ugly.