Search Results: "beh"

2 February 2025

Dave Hibberd: SOTA Trip Reports: Feb 02, 2025 - Bennachie

To Quote @MM0EFI and the GM0ESS gang, today was a particularly Amateur showing! Having spent all weekend locked in the curling rink ruining my knees and inflicting mild liver damage in the Aberdeen City Open competition, I needed some outside time away from people to stretch the legs and loosen my knees. With my teammates/guests shipped off early on account of our quality performance and the days fair drawin out now, I found myself with a free afternoon to have a quick run up something nearby before a 1640 sunset! Up the back of Bennachie is a quick steady ascent and in 13 years of living up here I ve never summited the big hill! Now is as good a time as any. In SOTA terms, this hill is GM/ES-061. In Geographical terms, it s around 20 miles inland from Aberdeen city here. I ve been experimenting with these Aliexpress whips since the end of last year and the forecast wind was low enough to take one into the hills. I cut and terminated 8x 2.5m radials for an effective ground plane last week and wanted to try that against the flat ribbon that it came with. The ascent was pleasant enough, got to the summit in good time, and out came my Quansheng radio to get the GM/ES-Society on 2m. First my Nagoya whip - called CQ and heard nothing, with general poor reports in WhatsApp I opted to get the slim-g up my aliexpress fibreglass mast. In an amateur showing last week, I broke the tip of the mast on Cat Law helping 2M0HSK do his first activation due to the wind, and had forgotten this until I summited this week. Squeezing my antenna on was tough, and after many failed attempts to get it up (the mast kept collapsing as I was rushing and not getting the friction hold on each section correctly) and still not hearing anything at all, I changed location and tried again. In my new position, I received 2M0RVZ 4/4 at best, but he was hearing my 5/9. Similarly GM5ALX and GM4JXP were patiently receiving me loud and clear but I couldn t hear them at all. I fiddled with settings and decided the receive path of the Quansheng must be fried or sad somehow, but I don t yet have a full set of diagnostics run. I ll take my Anytone on the next hill and compare them against each other I think. I gave up and moved to HF, getting my whip and new radials into the ground: 295B58E1-BA43-4348-A4C7-0B1E013C4006_1_102_o 375x500 Quick to deploy which is what I was after. My new 5m of coax with a choke fitted attached to the radio and we were off to the races - A convenient thing of beauty when it s up: 33C35D56-F470-46BB-B31E-F66361504A1C_1_102_o 375x500 I ve made a single guy with a sotabeams top insulator to brace against wind if need be, but that didn t need to be used today. I hit tune, and the G90 spent ages clicking away. In fact, tuning to 14.074, I could only see the famed FT8 signals at S2. What could be wrong here? Was it my new radials? the whip has behaved before Minutes turned into tens of minutes playing with everything, and eventually I worked out what was up - my coax only passed signal when I the PL259 connector at the antenna juuuust right. Once I did that, I could take the tuner out the system and work 20 spectacularly well. Until now, I d been tuning the coax only. Another Quality Hibby Build Job . That s what s wrong! I managed to struggle my way through a touch of QRM and my wonky cable woes to make enough contacts with some very patient chasers and a summit to summit before my frustration at the situation won out, and down the hill I went after a quick pack up period. I managed to beat the sunset - I think if the system had worked fine, I d have stayed on the hill for sunset. I think it s time for a new mast and a coax retermination!

Joachim Breitner: Coding on my eInk Tablet

For many years I wished I had a setup that would allow me to work (that is, code) productively outside in the bright sun. It s winter right now, but when its summer again it s always a bit. this weekend I got closer to that goal. TL;DR: Using code-server on a beefy machine seems to be quite neat.
Passively lit coding Passively lit coding

Personal history Looking back at my own old blog entries I find one from 10 years ago describing how I bought a Kobo eBook reader with the intent of using it as an external monitor for my laptop. It seems that I got a proof-of-concept setup working, using VNC, but it was tedious to set up, and I never actually used that. I subsequently noticed that the eBook reader is rather useful to read eBooks, and it has been in heavy use for that every since. Four years ago I gave this old idea another shot and bought an Onyx BOOX Max Lumi. This is an A4-sized tablet running Android and had the very promising feature of an HDMI input. So hopefully I d attach it to my laptop and it just works . Turns out that this never worked as well as I hoped: Even if I set the resolution to exactly the tablet s screen s resolution I got blurry output, and it also drained the battery a lot, so I gave up on this. I subsequently noticed that the tablet is rather useful to take notes, and it has been in sporadic use for that. Going off on this tangent: I later learned that the HDMI input of this device appears to the system like a camera input, and I don t have to use Boox s monitor app but could other apps like FreeDCam as well. This somehow managed to fix the resolution issues, but the setup still wasn t as convenient to be used regularly. I also played around with pure terminal approaches, e.g. SSH ing into a system, but since my usual workflow was never purely text-based (I was at least used to using a window manager instead of a terminal multiplexer like screen or tmux) that never led anywhere either.

VSCode, working remotely Since these attempts I have started a new job working on the Lean theorem prover, and working on or with Lean basically means using VSCode. (There is a very good neovim plugin as well, but I m using VSCode nevertheless, if only to make sure I am dogfooding our default user experience). My colleagues have said good things about using VSCode with the remote SSH extension to work on a beefy machine, so I gave this a try now as well, and while it s not a complete game changer for me, it does make certain tasks (rebuilding everything after a switching branches, running the test suite) very convenient. And it s a bit spooky to run these work loads without the laptop s fan spinning up. In this setup, the workspace is remote, but VSCode still runs locally. But it made me wonder about my old goal of being able to work reasonably efficient on my eInk tablet. Can I replicate this setup there? VSCode itself doesn t run on Android directly. There are project that run a Linux chroot or in termux on the Android system, and then you can VNC to connect to it (e.g. on Andronix) but that did not seem promising. It seemed fiddly, and I probably should take it easy on the tablet s system.

code-server, running remotely A more promising option is code-server. This is a fork of VSCode (actually of VSCodium) that runs completely on the remote machine, and the client machine just needs a browser. I set that up this weekend and found that I was able to do a little bit of work reasonably.

Access With code-server one has to decide how to expose it safely enough. I decided against the tunnel-over-SSH option, as I expected that to be somewhat tedious to set up (both initially and for each session) on the android system, and I liked the idea of being able to use any device to work in my environment. I also decided against the more involved reverse proxy behind proper hostname with SSL setups, because they involve a few extra steps, and some of them I cannot do as I do not have root access on the shared beefy machine I wanted to use. That left me with the option of using a code-server s built-in support for self-signed certificates and a password:
$ cat .config/code-server/config.yaml
bind-addr: 1.2.3.4:8080
auth: password
password: xxxxxxxxxxxxxxxxxxxxxxxx
cert: true
With trust-on-first-use this seems reasonably secure.

Service To keep code-server running I created a systemd service that s managed by my user s systemd instance:
~ $ cat ~/.config/systemd/user/code-server.service
[Unit]
Description=code-server
After=network-online.target
[Service]
Environment=PATH=/home/joachim/.nix-profile/bin:/nix/var/nix/profiles/default/bin:/usr/local/bin:/usr/bin:/bin:/usr/local/games:/usr/games
ExecStart=/nix/var/nix/profiles/default/bin/nix run nixpkgs#code-server
[Install]
WantedBy=default.target
(I am using nix as a package manager on a Debian system there, hence the additional PATH and complex ExecStart. If you have a more conventional setup then you do not have to worry about Environment and can likely use ExecStart=code-server. For this to survive me logging out I had to ask the system administrator to run loginctl enable-linger joachim, so that systemd allows my jobs to linger.

Git credentials The next issue to be solved was how to access the git repositories. The work is all on public repositories, but I still need a way to push my work. With the classic VSCode-SSH-remote setup from my laptop, this is no problem: My local SSH key is forwarded using the SSH agent, so I can seamlessly use that on the other side. But with code-server there is no SSH key involved. I could create a new SSH key and store it on the server. That did not seem appealing, though, because SSH keys on Github always have full access. It wouldn t be horrible, but I still wondered if I can do better. I thought of creating fine-grained personal access tokens that only me to push code to specific repositories, and nothing else, and just store them permanently on the remote server. Still a neat and convenient option, but creating PATs for our org requires approval and I didn t want to bother anyone on the weekend. So I am experimenting with Github s git-credential-manager now. I have configured it to use git s credential cache with an elevated timeout, so that once I log in, I don t have to again for one workday.
$ nix-env -iA nixpkgs.git-credential-manager
$ git-credential-manager configure
$ git config --global credential.credentialStore cache
$ git config --global credential.cacheOptions "--timeout 36000"
To login, I have to https://github.com/login/device on an authenticated device (e.g. my phone) and enter a 8-character code. Not too shabby in terms of security. I only wish that webpage would not require me to press Tab after each character This still grants rather broad permissions to the code-server, but at least only temporarily

Android setup On the client side I could now open https://host.example.com:8080 in Firefox on my eInk Android tablet, click through the warning about self-signed certificates, log in with the fixed password mentioned above, and start working! I switched to a theme that supposedly is eInk-optimized (eInk by Mufanza). It s not perfect (e.g. git diffs are unhelpful because it is not possible to distinguish deleted from added lines), but it s a start. There are more eInk themes on the official Visual Studio Marketplace, but because code-server is a fork it cannot use that marketplace, and for example this theme isn t on Open-VSX. For some reason the F11 key doesn t work, but going fullscreen is crucial, because screen estate is scarce in this setup. I can go fullscreen using VSCode s command palette (Ctrl-P) and invoking the command there, but Firefox often jumps out of the fullscreen mode, which is annoying. I still have to pay attention to when that s happening; maybe its the Esc key, which I am of course using a lot due to me using vim bindings. A more annoying problem was that on my Boox tablet, sometimes the on-screen keyboard would pop up, which is seriously annoying! It took me a while to track this down: The Boox has two virtual keyboards installed: The usual Google ASOP keyboard, and the Onyx Keyboard. The former is clever enough to stay hidden when there is a physical keyboard attached, but the latter isn t. Moreover, pressing Shift-Ctrl on the physical keyboard rotates through the virtual keyboards. Now, VSCode has many keyboard shortcuts that require Shift-Ctrl (especially on an eInk device, where you really want to avoid using the mouse). And the limited settings exposed by the Boox Android system do not allow you configure that or disable the Onyx keyboard! To solve this, I had to install the KISS Launcher, which would allow me to see more Android settings, and in particular allow me to disable the Onyx keyboard. So this is fixed. I was hoping to improve the experience even more by opening the web page as a Progressive Web App (PWA), as described in the code-server FAQ. Unfortunately, that did not work. Firefox on Android did not recognize the site as a PWA (even though it recognizes a PWA test page). And I couldn t use Chrome either because (unlike Firefox) it would not consider a site with a self-signed certificate as a secure context, and then code-server does not work fully. Maybe this is just some bug that gets fixed in later versions. I did not work enough with this yet to assess how much the smaller screen estate, the lack of colors and the slower refresh rate will bother me. I probably need to hide Lean s InfoView more often, and maybe use the Error Lens extension, to avoid having to split my screen vertically. I also cannot easily work on a park bench this way, with a tablet and a separate external keyboard. I d need at least a table, or some additional piece of hardware that turns tablet + keyboard into some laptop-like structure that I can put on my, well, lap. There are cases for Onyx products that include a keyboard, and maybe they work on the lap, but they don t have the Trackpoint that I have on my ThinkPad TrackPoint Keyboard II, and how can you live without that?

Conclusion After this initial setup chances are good that entering and using this environment is convenient enough for me to actually use it; we will see when it gets warmer. A few bits could be better. In particular logging in and authenticating GitHub access could be both more convenient and more safe I could imagine that when I open the page I confirm that on my phone (maybe with a fingerprint), and that temporarily grants access to the code-server and to specific GitHub repositories only. Is that easily possible?

31 January 2025

Gunnar Wolf: ChatGPT is bullshit

This post is an unpublished review for ChatGPT is bullshit
As people around the world understand how LLMs behave, more and more people wonder as to why these models hallucinate, and what can be done about to reduce it. This provocatively named article by Michael Townsen Hicks, James Humphries and Joe Slater bring is an excellent primer to better understanding how LLMs work and what to expect from them. As humans carrying out our relations using our language as the main tool, we are easily at awe with the apparent ease with which ChatGPT (the first widely available, and to this day probably the best known, LLM-based automated chatbot) simulates human-like understanding and how it helps us to easily carry out even daunting data aggregation tasks. It is common that people ask ChatGPT for an answer and, if it gets part of the answer wrong, they justify it by stating that it s just a hallucination. Townsen et al. invite us to switch from that characterization to a more correct one: LLMs are bullshitting. This term is formally presented by Frankfurt [1]. To Bullshit is not the same as to lie, because lying requires to know (and want to cover) the truth. A bullshitter not necessarily knows the truth, they just have to provide a compelling description, regardless of what is really aligned with truth. After introducing Frankfurt s ideas, the authors explain the fundamental ideas behind LLM-based chatbots such as ChatGPT; a Generative Pre-trained Transformer (GPT) s have as their only goal to produce human-like text, and it is carried out mainly by presenting output that matches the input s high-dimensional abstract vector representation, and probabilistically outputs the next token (word) iteratively with the text produced so far. Clearly, a GPT s ask is not to seek truth or to convey useful information they are built to provide a normal-seeming response to the prompts provided by their user. Core data are not queried to find optimal solutions for the user s requests, but are generated on the requested topic, attempting to mimic the style of document set it was trained with. Erroneous data emitted by a LLM is, thus, not equiparable with what a person could hallucinate with, but appears because the model has no understanding of truth; in a way, this is very fitting with the current state of the world, a time often termed as the age of post-truth [2]. Requesting an LLM to provide truth in its answers is basically impossible, given the difference between intelligence and consciousness: Following Harari s definitions [3], LLM systems, or any AI-based system, can be seen as intelligent, as they have the ability to attain goals in various, flexible ways, but they cannot be seen as conscious, as they have no ability to experience subjectivity. This is, the LLM is, by definition, bullshitting its way towards an answer: their goal is to provide an answer, not to interpret the world in a trustworthy way. The authors close their article with a plea for literature on the topic to adopt the more correct bullshit term instead of the vacuous, anthropomorphizing hallucination . Of course, being the word already loaded with a negative meaning, it is an unlikely request. This is a great article that mixes together Computer Science and Philosophy, and can shed some light on a topic that is hard to grasp for many users. [1] Frankfurt, Harry (2005). On Bullshit. Princeton University Press. [2] Zoglauer, Thomas (2023). Constructed truths: truth and knowledge in a post-truth world. Springer. [3] Harari, Yuval Noah (2023. Nexus: A Brief History of Information Networks From the Stone Age to AI. Random House.

26 January 2025

Russ Allbery: Review: Dark Matters

Review: Dark Matters, by Michelle Diener
Series: Class 5 #4
Publisher: Eclipse
Copyright: October 2019
ISBN: 0-6454658-6-0
Format: Kindle
Pages: 307
Dark Matters is the fourth book in the science fiction semi-romance Class 5 series. There are spoilers for all of the previous books, and although enough is explained that you could make sense of the story starting here, I wouldn't recommend it. As with the other books in the series, it follows new protagonists, but the previous protagonists make an appearance. You will be unsurprised to hear that the Tecran kidnapped yet another Earth woman. The repetitiveness of the setup would be more annoying if the book took itself too seriously, but it doesn't, and so I mostly find it entertaining. I thought Diener was going to dodge the obvious series structure, but now I am wondering if we're going to end up with one woman per Class 5 ship after all. Lucy is not on a ship, however, Tecran or otherwise. She is a captive in a military research facility on the Tecran home world. The Tecran are in very deep trouble given the events of the previous book and have decided that Lucy's existence is a liability. Only the intervention of some sympathetic Tecran scientists she partly befriended during her captivity lets her escape the facility before it's destroyed. Now she's alone, on an alien world, being hunted by the military. It's not entirely the fault of this book that it didn't tell the story that I wanted to read. The setup for Dark Matters implies this book will see the arrival of consequences for the Tecran's blatant violations of the Sentient Beings Agreement. I was looking forward to a more political novel about how such consequences could be administered. This is the sort of problem that we struggle with in our politics: Collective punishment isn't acceptable, but there have to be consequences sufficient to ensure that a state doesn't repeat the outlawed behavior, and yet attempting to deliver those consequences feels like occupation and can set off worse social ruptures and even atrocities. I wasn't expecting that deep of political analysis of what is, after all, a lighthearted SF adventure series, but Diener has been willing to touch on hard problems. The ethics of violence has been an ongoing theme of the series. Alas for me, this is not what we get. The arriving cavalry, in the form of a Class 5 and the inevitable Grih hunk to serve as the love interest du jour, quickly become more interested in helping Lucy elude pursuers (or escape captors) than in the delicate political situation. The conflict between the local population is a significant story element, but only as backdrop. Instead, this reads like a thriller or an action movie, complete with alien predators and a cinematic set piece finale. The political conflict between the Tecran and the United Council does reach a conclusion of sorts, but it's not that satisfying. Perhaps some of the political fallout will happen in future books, but here Diener simplifies the morality of the story in the climax and dodges out of the tricky ethical and social challenge of how to punish a sovereign nation. One of the things I like about this series is that it takes moral indignation seriously, but now that Diener has raised the (correct) complication that people have strong motivations to find excuses for the actions of their own side, I hope she can find a believable political resolution that isn't simple brute force. This entry in the series wasn't bad, but it didn't grab me. Lucy was fine as a protagonist; her ability to manipulate the Tecran into making mistakes fits the longer time she's had to study them and keeps her distinct from the other protagonists. But the small bit of politics we do see is unsatisfying and conveniently simplistic, and this book mostly degenerates into generic action sequences. Bane, the Class 5 ship featured in this story, is great when he's active, and I continue to be entertained by the obsession the Class 5 ships have with Earth women, but he's sidelined for too much of the story. I felt like Diener focused on the least interesting part of the story setup. If you've read this far, there's nothing wrong with this entry. You'll probably want to keep reading. But it felt like a missed opportunity. Followed in publication order by Dark Ambitions, a novella that returns to Rose to tell a side story. The next novel is Dark Class, in which we'll presumably see the last kidnapped Earth woman. Rating: 6 out of 10

25 January 2025

Steve Kemp: The CP/M emulator now works better!

I keep saying I'm "done" with my CP/M emulator, but then I keep overhauling it in significant ways. Today is no exception. In the past the emulator used breakpoints to detect when calls to the system BIOS, or BDOS, were made. That was possible because the BIOS and BDOS entry points are at predictable locations. For example a well-behaved program might make a system call with code like this:
    LD A,42
    LD C,4
    CALL 0x0005
So setting a breakpoint on 0x0005 would let you detect a system-call was being made, inspect the registers to see which system-call was being made and then carry out the appropriate action in your emulator before returning control back to the program. Unfortunately some binaries patch the RAM, changing the contents of the entry points, or changing internal jump-tables, etc. The end result is that sometimes code running at the fixed addresses is not your BIOS at all, but something else. By trapping/faulting/catching execution here you break things, badly. So today's new release fixes that! No more breakpoints. Instead we deploy a "real BDOS" in RAM that will route system-calls to our host emulator via a clever trick. For BDOS functions the C-register will contain the system call to operate, our complete BDOS implementation is:
    OUT (C),C
    RET
The host program can catch writes to output ports, and will know that "OUT (3), 3" means "Invoke system call #3", for example. This means binary patches to entry-points, or any internal jump-tables won't confuse things and so long as control eventually reaches my BIOS or BDOS code areas things will work. I also added a new console-input driver, since I have a notion of pluggable input and output devices, which just reads input from a file. Now I can prove that my code works. Pass the following file to the input-driver and we have automated testing:
A:
ERA HELLO.COM
ERA HELLO.HEX
ERA HELLO.PRN
hello
ASM HELLO
LOAD HELLO
DDT HELLO.com
t
t
t
t
t
t
t
t
t
C-c
EXIT
Here we: I can test the output and confirm there are no regressions. Neat. Anyway new release, today. Happy.

20 January 2025

Divine Attah-Ohiemi: Progress Report: First Half of My Outreachy Internship

Hello everyone!, I m excited to share a progress report on my Outreachy internship with the Debian community. As I reach the halfway point of this journey, I want to reflect on what I ve accomplished so far and outline my modified goals for the second half of the internship. In truth, there wasn t a strict timeline for my project migrating Debian webpage content to Hugo because the original repository contained thousands of pages. The initial goal was to develop a proof of concept for: Thanks to our daily standups, where we brainstorm and revise contributions, we ve made significant progress. The wiki documentation discussing the technical decisions taken to meet these goals is currently in progress here. During the first half of my internship, I have improved and refined my skills in several areas. I learned new Markdown syntaxes, studied and utilized Apache's mod_rewrite, and halfway studied GNU Make to use Perl scripts for processing data for dynamic content. I recommend Managing Projects with GNU Make by Robert Mecklenburg it's a great book for beginners! While I didn t get stuck on any particular goal, the most challenging aspect was adding Hugo aliases to help with Apache's multilingual content negotiation. The way the webwml repository generates multilingual content differs from debianhugo. For instance, in webwml, the structure looks like this: english/index.wml -> /index.en.html (with a symlink from index.html to index.en.html) and french/index.wml -> /index.fr.html. In contrast, debianhugo uses en/_index.md -> /index.html and fr/_index.md -> /fr/index.html. Apache's multilingual content negotiation checks for index.<user preferred lang code>.html in the current directory, which works well with webwml since all related translations are generated in the same directory. However, with debianhugo using subdirectories for languages other than English, we had to set up aliases for every other language page to be generated in the frontmatter. For example, in fr/_index.md, we added this to the front matter:
...
aliases:
  - /index.fr.html
...
This setup allows Hugo to generate multilingual HTML files in the initial home directory solely for the purpose of setting up a 301 redirect to the same page in the language subdirectory. However, if the client sets their preferred language to English, Apache content negotiation tries to find /index.en.html. If it doesn t find it, it defaults to any other language-suffixed file, which can lead to unexpected behavior. For example, if English is set as the preferred language, accessing the site may serve /index.fr.html, which then redirects to /fr/index.html. This was a significant challenge, and you can see a demo of this hosted here. If I were to start the project over, I would document every decision as I make them in the wiki, no matter how rough the documentation turns out. Waiting until the midpoint of the project to document was not a good idea. As I move into the second half of my internship, the goals we ve set include improving our project wiki documentation and continuing the migration process while enhancing the user experience of complicated sections. I m looking forward to making even more progress and sharing my journey with you all. Happy coding!

10 January 2025

Sergio Talens-Oliag: Testing New User Tools

On recent weeks I ve had some time to scratch my own itch on matters related to tools I use daily on my computer, namely the desktop / window manager and my text editor of choice. This post is a summary of what I tried, how it worked out and my short and medium-term plans related to them.

Desktop / WMOn the desktop / window manager front I ve been using Cinnamon on Debian and Ubuntu systems since Gnome 3 was published (I never liked version 3, so I decided to move to something similar to Gnome 2, including the keyboard shortcuts). In fact I ve never been a fan of Desktop environments, before Gnome I used OpenBox and IceWM because they where a lot faster than desktop systems on my hardware at the time and I was using them only to place one or two windows on multiple workspaces using mainly the keyboard for my interactions (well, except for the web browsers and the image manipulation programs). Although I was comfortable using Cinnamon, some years ago I tried to move to i3, a tilling window manager for X11 that looked like a good choice for me, but I didn t have much time to play with it and never used it enough to make me productive with it (I didn t prepare a complete configuration nor had enough time to learn the new shortcuts, so I went back to Cinnamon and never tried again). Anyway, some weeks ago I updated my work machine OS (it was using Ubuntu 22.04 LTS and I updated it to the 24.04 LTS version) and the Cinnamon systray applet stopped working as it used to do (in fact I still have to restart Cinnamon after starting a session to make it work) and, as I had some time, I decided to try a tilling window manager again, but now I decided to go for SwayWM, as it uses Wayland instead of X11.

Sway configurationOn my ~/.config/sway/config I tuned some things:
  • Set fuzzel as the application launcher.
  • Installed manually the shikane application and created a configuration to be executed always when sway is started / reloaded (I adjusted my configuration with wdisplays and used shikanectl to save it).
  • Added support for running the xdg-desktop-portal-wlr service.
  • Enabled the swayidle command to lock the screen after some time of inactivity.
  • Adjusted the keyboard to use the es key map
  • Added some keybindings to make my life easier, including the use of grimm and swappy to take screenshots
  • Configured waybar as the environment bar.
  • Added a shell script to start applications when sway is started (it uses swaymsg to execute background commands and the i3toolwait script to wait for the
    #!/bin/sh
    # VARIABLES
    CHROMIUM_LOCAL_STATE="$HOME/.config/google-chrome/Local State"
    I3_TOOLWAIT="$HOME/.config/sway/scripts/i3-toolwait"
    # Functions
    chromium_profile_dir()  
      jq -r ".profile.info_cache to_entries map( (.value.name): .key ) add .\"$1\" // \"\"" "$CHROMIUM_LOCAL_STATE"
     
    # MAIN
    IGZ_PROFILE_DIR="$(chromium_profile_dir "sergio.talens@intelygenz.com")"
    OURO_PROFILE_DIR="$(chromium_profile_dir "sergio.talens@nxr.global")"
    PERSONAL_PROFILE_DIR="$(chromium_profile_dir "stalens@gmail.com")"
    # Common programs
    swaymsg "exec nextcloud --background"
    swaymsg "exec nm-applet"
    # Run spotify on the first workspace (it is mapped to the laptop screen)
    swaymsg -q "workspace 1"
    $ I3_TOOLWAIT  "spotify"
    # Run tmux on the
    swaymsg -q "workspace 2"
    $ I3_TOOLWAIT  -- foot tmux a -dt sto
    wp_num="3"
    if [ "$OURO_PROFILE_DIR" ]; then
      swaymsg -q "workspace $wp_num"
      $ I3_TOOLWAIT  -m ouro-browser -- google-chrome --profile-directory="$OURO_PROFILE_DIR"
      wp_num="$((wp_num+1))"
    fi
    if [ "$IGZ_PROFILE_DIR" ]; then
      swaymsg -q "workspace $wp_num"
      $ I3_TOOLWAIT  -m igz-browser -- google-chrome --profile-directory="$IGZ_PROFILE_DIR"
      wp_num="$((wp_num+1))"
    fi
    if [ "$PERSONAL_PROFILE_DIR" ]; then
      swaymsg -q "workspace $wp_num"
      $ I3_TOOLWAIT  -m personal-browser -- google-chrome --profile-directory="$PERSONAL_PROFILE_DIR"
      wp_num="$((wp_num+1))"
    fi
    # Open the browser without setting the profile directory if none was found
    if [ "$wp_num" = "3" ]; then
      swaymsg -q "workspace $wp_num"
      $ I3_TOOLWAIT  google-chrome
      wp_num="$((wp_num+1))"
    fi
    swaymsg -q "workspace $wp_num"
    $ I3_TOOLWAIT  evolution
    wp_num="$((wp_num+1))"
    swaymsg -q "workspace $wp_num"
    $ I3_TOOLWAIT  slack
    wp_num="$((wp_num+1))"
    # Open a private browser and a console in the last workspace
    swaymsg -q "workspace $wp_num"
    $ I3_TOOLWAIT  -- google-chrome --incognito
    $ I3_TOOLWAIT  foot
    # Go back to the second workspace for keepassxc
    swaymsg "workspace 2"
    $ I3_TOOLWAIT  keepassxc

ConclusionAfter using Sway for some days I can confirm that it is a good choice for me, but some of the components needed to make it work as I want are too new and not available on the Ubuntu 24.04 LTS repositories, so I decided to go back to Cinnamon and try Sway again in the future, although I added more workspaces to my setup (now they are only available on the main monitor, the laptop screen is fixed while there is a big monitor connected), added some additional keyboard shortcuts and installed or updated some applets.

Text editorWhen I started using Linux many years ago I used vi/vim and emacs as my text editors (vi for plain text and emacs for programming and editing HTML/XML), but eventually I moved to vim as my main text editor and I ve been using it since (well, I moved to neovim some time ago, although I kept my old vim configuration). To be fair I m not as expert as I could be with vim, but I m productive with it and it has many plugins that make my life easier on my machines, while keeping my ability to edit text and configurations on any system that has a vi compatible editor installed. For work reasons I tried to use Visual Studio Code last year, but I ve never really liked it and almost everything I do with it I can do with neovim (i. e. I even use copilot with it). Besides, I m a heavy terminal user (I use tmux locally and via ssh) and I like to be able to use my text editor on my shell sessions, and code does not work like that. The only annoying thing about vim/neovim is its configuration (well, the problem is that I have a very old one and probably should spend some time fixing and updating it), but, as I said, it s been working well for me for a long time, so I never really had the motivation to do it. Anyway, after finishing my desktop tests I saw that I had the Helix editor installed for some time but I never tried it, so I decided to give it a try and see if it could be a good replacement for neovim on my environments (the only drawback is that as it is not vi compatible, I would need to switch back to vi mode when working on remote systems, but I guess I could live with that). I ran the helix tutorial and I liked it, so I decided to configure and install the Language Servers I can probably take advantage of on my daily work on my personal and work machines and see how it works.

Language server installationsA lot of manual installations are needed to get the language servers working what I did on my machines is more or less the following:
# AWK
sudo npm i -g 'awk-language-server@>=0.5.2'
# BASH
sudo apt-get install shellcheck shfmt
sudo npm i -g bash-language-server
# C/C++
sudo apt-get install clangd
# CSS, HTML, ESLint, JSON, SCS
sudo npm i -g vscode-langservers-extracted
# Docker
sudo npm install -g dockerfile-language-server-nodejs
# Docker compose
sudo npm install -g @microsoft/compose-language-service
# Helm
app="helm_ls_linux_amd64"
url="$(
  curl -s https://api.github.com/repos/mrjosh/helm-ls/releases/latest  
    jq -r ".assets[]   select(.name == \"$app\")   .browser_download_url"
)"
curl -L "$url" --output /tmp/helm_ls
sudo install /tmp/helm_ls /usr/local/bin
rm /tmp/helm_ls
# Markdown
app="marksman-linux-x64"
url="$(
  curl -s https://api.github.com/repos/artempyanykh/marksman/releases/latest  
    jq -r ".assets[]   select(.name == \"$app\")   .browser_download_url"
)"
curl -L "$url" --output /tmp/marksman
sudo install /tmp/marksman /usr/local/bin
rm /tmp/marksman
# Python
sudo npm i -g pyright
# Rust
rustup component add rust-analyzer
# SQL
sudo npm i -g sql-language-server
# Terraform
sudo apt-get install terraform-ls
# TOML
cargo install taplo-cli --locked --features lsp
# YAML
sudo npm install --global yaml-language-server
# JavaScript, TypeScript
sudo npm install -g typescript-language-server typescript
sudo npm install -g --save-dev --save-exact @biomejs/biome

Helix configurationThe helix configuration is done on a couple of toml files that are placed on the ~/.config/helix directory, the config.toml file I used is this one:
theme = "solarized_light"
[editor]
line-number = "relative"
mouse = false
[editor.statusline]
left = ["mode", "spinner"]
center = ["file-name"]
right = ["diagnostics", "selections", "position", "file-encoding", "file-line-ending", "file-type"]
separator = " "
mode.normal = "NORMAL"
mode.insert = "INSERT"
mode.select = "SELECT"
[editor.cursor-shape]
insert = "bar"
normal = "block"
select = "underline"
[editor.file-picker]
hidden = false
[editor.whitespace]
render = "all"
[editor.indent-guides]
render = true
character = " " # Some characters that work well: " ", " ", " ", " "
skip-levels = 1
And to configure the language servers I used the following language-servers.toml file:
[[language]]
name = "go"
auto-format = true
formatter =   command = "goimports"  
[[language]]
name = "javascript"
language-servers = [
  "typescript-language-server", # optional
  "vscode-eslint-language-server",
]
[language-server.rust-analyzer.config.check]
command = "clippy"
[language-server.sql-language-server]
command = "sql-language-server"
args = ["up", "--method", "stdio"]
[[language]]
name = "sql"
language-servers = [ "sql-language-server" ]
[[language]]
name = "hcl"
language-servers = [ "terraform-ls" ]
language-id = "terraform"
[[language]]
name = "tfvars"
language-servers = [ "terraform-ls" ]
language-id = "terraform-vars"
[language-server.terraform-ls]
command = "terraform-ls"
args = ["serve"]
[[language]]
name = "toml"
formatter =   command = "taplo", args = ["fmt", "-"]  
[[language]]
name = "typescript"
language-servers = [
  "typescript-language-server",
  "vscode-eslint-language-server",
]

Neovim configurationAfter a little while I noticed that I was going to need some time to get used to helix and the most interesting thing for me was the easy configuration and the language server integrations, but as I am already comfortable with neovim and just had installed the language server support tools on my machines I just need to configure them for neovim and I can keep using it for a while. As I said my configuration is old, to configure neovim I have the following init.vim file on my ~/.config/nvim folder:
set runtimepath^=~/.vim runtimepath+=~/.vim/after
let &packpath=&runtimepath
source ~/.vim/vimrc
" load lua configuration
lua require('config')
With that configuration I keep my old vimrc (it is a little bit messy, but it works) and I use a lua configuration file for the language servers and some additional neovim plugins on the ~/.config/nvim/lua/config.lua file:
-- -----------------------
-- BEG: LSP Configurations
-- -----------------------
-- AWS (awk_ls)
require'lspconfig'.awk_ls.setup 
-- Bash (bashls)
require'lspconfig'.bashls.setup 
-- C/C++ (clangd)
require'lspconfig'.clangd.setup 
-- CSS (cssls)
require'lspconfig'.cssls.setup 
-- Docker (dockerls)
require'lspconfig'.dockerls.setup 
-- Docker Compose
require'lspconfig'.docker_compose_language_service.setup 
-- Golang (gopls)
require'lspconfig'.gopls.setup 
-- Helm (helm_ls)
require'lspconfig'.helm_ls.setup 
-- Markdown
require'lspconfig'.marksman.setup 
-- Python (pyright)
require'lspconfig'.pyright.setup 
-- Rust (rust-analyzer)
require'lspconfig'.rust_analyzer.setup 
-- SQL (sqlls)
require'lspconfig'.sqlls.setup 
-- Terraform (terraformls)
require'lspconfig'.terraformls.setup 
-- TOML (taplo)
require'lspconfig'.taplo.setup 
-- Typescript (ts_ls)
require'lspconfig'.ts_ls.setup 
-- YAML (yamlls)
require'lspconfig'.yamlls.setup 
  settings =  
    yaml =  
      customTags =   "!reference sequence"  
     
   
 
-- -----------------------
-- END: LSP Configurations
-- -----------------------
-- ---------------------------------
-- BEG: Autocompletion configuration
-- ---------------------------------
-- Ref: https://github.com/neovim/nvim-lspconfig/wiki/Autocompletion
--
-- Pre requisites:
--
--   # Packer
--   git clone --depth 1 https://github.com/wbthomason/packer.nvim \
--      ~/.local/share/nvim/site/pack/packer/start/packer.nvim
--
--   # Start nvim and run :PackerSync or :PackerUpdate
-- ---------------------------------
local use = require('packer').use
require('packer').startup(function()
  use 'wbthomason/packer.nvim' -- Packer, useful to avoid removing it with PackerSync / PackerUpdate
  use 'neovim/nvim-lspconfig' -- Collection of configurations for built-in LSP client
  use 'hrsh7th/nvim-cmp' -- Autocompletion plugin
  use 'hrsh7th/cmp-nvim-lsp' -- LSP source for nvim-cmp
  use 'saadparwaiz1/cmp_luasnip' -- Snippets source for nvim-cmp
  use 'L3MON4D3/LuaSnip' -- Snippets plugin
end)
-- Add additional capabilities supported by nvim-cmp
local capabilities = require("cmp_nvim_lsp").default_capabilities()
local lspconfig = require('lspconfig')
-- Enable some language servers with the additional completion capabilities offered by nvim-cmp
local servers =   'clangd', 'rust_analyzer', 'pyright', 'ts_ls'  
for _, lsp in ipairs(servers) do
  lspconfig[lsp].setup  
    -- on_attach = my_custom_on_attach,
    capabilities = capabilities,
   
end
-- luasnip setup
local luasnip = require 'luasnip'
-- nvim-cmp setup
local cmp = require 'cmp'
cmp.setup  
  snippet =  
    expand = function(args)
      luasnip.lsp_expand(args.body)
    end,
   ,
  mapping = cmp.mapping.preset.insert( 
    ['<C-u>'] = cmp.mapping.scroll_docs(-4), -- Up
    ['<C-d>'] = cmp.mapping.scroll_docs(4), -- Down
    -- C-b (back) C-f (forward) for snippet placeholder navigation.
    ['<C-Space>'] = cmp.mapping.complete(),
    ['<CR>'] = cmp.mapping.confirm  
      behavior = cmp.ConfirmBehavior.Replace,
      select = true,
     ,
    ['<Tab>'] = cmp.mapping(function(fallback)
      if cmp.visible() then
        cmp.select_next_item()
      elseif luasnip.expand_or_jumpable() then
        luasnip.expand_or_jump()
      else
        fallback()
      end
    end,   'i', 's'  ),
    ['<S-Tab>'] = cmp.mapping(function(fallback)
      if cmp.visible() then
        cmp.select_prev_item()
      elseif luasnip.jumpable(-1) then
        luasnip.jump(-1)
      else
        fallback()
      end
    end,   'i', 's'  ),
   ),
  sources =  
      name = 'nvim_lsp'  ,
      name = 'luasnip'  ,
   ,
 
-- ---------------------------------
-- END: Autocompletion configuration
-- ---------------------------------

ConclusionI guess I ll keep helix installed and try it again on some of my personal projects to see if I can get used to it, but for now I ll stay with neovim as my main text editor and learn the shortcuts to use it with the language servers.

8 January 2025

Dirk Eddelbuettel: RcppGetconf 0.0.4 on CRAN: Updates

A minor package update, the first in over six years, for the RcppGetconf package for reading system configuration not unlike getconf from the libc library is now on CRAN The changes are all minor package maintenance items of keeping URLs, continuous integration, and best practices current. We had two helper scripts use bash in their shebangs, and we just got dinged in one of them. Tedious as this can at times seem, it ensures CRAN packages do in fact compile just about anywhere which is a Good Thing (TM) so we obliged and updated the package with that change and all the others that had accumulated over six years. No interface or behaviour changes, just maintenance as one does at times. The short list of changes in this release follows:

Changes in inline version 0.0.4 (2025-01-07)
  • Dynamically linked compiled code is now registered in NAMESPACE
  • The continuous integration setup was update several times
  • The README was updated with current badges and URLs
  • The DESCRIPTION file now uses Authors@R
  • The configure and cleanup scripts use /bin/sh

Courtesy of my CRANberries, there is also a diffstat report of changes relative to the previous release. More about the package is at the local RcppGetconf page and the GitHub repo.

This post by Dirk Eddelbuettel originated on his Thinking inside the box blog. If you like this or other open-source work I do, you can now sponsor me at GitHub.

2 January 2025

Colin Watson: Free software activity in December 2024

Most of my Debian contributions this month were sponsored by Freexian, as well as one direct donation via Liberapay (thanks!). OpenSSH I issued a bookworm update with a number of fixes that had accumulated over the last year, especially fixing GSS-API key exchange which was quite broken in bookworm. base-passwd A few months ago, the adduser maintainer started a discussion with me (as the base-passwd maintainer) and the shadow maintainer about bringing all three source packages under one team, since they often need to cooperate on things like user and group names. I agreed, but hadn t got round to doing anything about it until recently. I ve now officially moved it under team maintenance. debconf Gioele Barabucci has been working on eliminating duplicated code between debconf and cdebconf, ultimately with the goal of migrating to cdebconf (which I m not sure I m convinced of as a goal, but if we can make improvements to both packages as part of working towards it then there s no harm in that). I finally got round to reviewing and merging confmodule changes in each of debconf and cdebconf. This caused an installer regression due to a weirdness in cdebconf-udeb s packaging, which I fixed - sorry about that! I ve also been dealing with a few patch submissions that had been in my queue for a long time, but more on that next month if all goes well. CI issues I noticed and fixed a problem with Restrictions: needs-sudo in autopkgtest. I fixed broken aptly images in the Salsa CI pipeline. Python team Last month, I mentioned some progress on sorting out the multipart vs. python-multipart name conflict in Debian (#1085728), and said that I thought we d be able to finish it soon. I was right! We got it all done this month: The Python 3.13 transition continues, and last month we were able to add it to the supported Python versions in testing. (The next step will be to make it the default.) I fixed lots of problems in aid of this, including: Sphinx 8.0 removed some old intersphinx_mapping syntax which turned out to still be in use by many packages in Debian. The fixes for this were individually trivial, but there were a lot of them: I found that twisted 24.11.0 broke tests in buildbot and wokkel, and fixed those. I packaged python-flatdict, needed for a new upstream version of python-semantic-release. I tracked down a test failure in vdirsyncer (which I ve been using for some years, but had never previously needed to modify) and contributed a fix upstream. I fixed some packages to tolerate future versions of dh-python that will drop their dependency on python3-setuptools: I fixed django-cte to remove a build-dependency on the obsolete python3-nose package. I added Django 5.1 support to django-polymorphic. (There are a number of other packages that still need work here.) I fixed various other build/test failures: I upgraded these packages to new upstream versions: I updated the team s library style guide to remove material related to Python 2 and early versions of Python 3, which is no longer relevant to any current Python packaging work. Other Python upstream work I happened to notice a Twisted upstream issue requesting the removal of the deprecated twisted.internet.defer.returnValue, realized it was still used in many places in Debian, and went on a PR-filing spree informed by codesearch to try to reduce the future impact of such a change on Debian: Other small fixes Santiago Vila has been building the archive with make --shuffle (also see its author s explanation). I fixed associated bugs in cccc (contributed upstream), groff, and spectemu. I backported an upstream patch to putty to fix undefined behaviour that affected use of the small keypad . I removed groff s Recommends: libpaper1 (#1091375, #1091376), since it isn t currently all that useful and was getting in the way of a transition to libpaper2. I filed an upstream bug suggesting better integration in this area.

31 December 2024

Russ Allbery: Review: Metal from Heaven

Review: Metal from Heaven, by August Clarke
Publisher: Erewhon
Copyright: November 2024
ISBN: 1-64566-099-0
Format: Kindle
Pages: 443
Metal from Heaven is industrial-era secondary-world fantasy with a literary bent. It is a complete story in one book, and I would be very surprised by a sequel. Clarke previously wrote the Scapegracers young-adult trilogy, which got excellent reviews and a few award nominations, as H.A. Clarke. This is his first adult novel.
Know I adore you. Look out over the glow. The cities sundered, their machines inverted, mountains split and prairies blazing, that long foreseen Hereafter crowning fast. This calamity is a promise made to you. A prayer to you, and to your shadow which has become my second self, tucked behind my eye and growing in tandem with me, pressing outwards through the pupil, the smarter, truer, almost bursting reason for our wrath. Do not doubt me. Just look. Watch us rise as the sun comes up over the beauty. The future stains the bleakness so pink. When my violence subsides, we will have nothing, and be champions.
Marney Honeycutt is twelve years old, a factory worker, and lustertouched. She works in the Yann I. Chauncey Ichorite Foundry in Ignavia City, alongside her family and her best friend, shaping the magical metal ichorite into the valuable industrial products of a new age of commerce and industry. She is the oldest of the lustertouched, the children born to factory workers and poisoned by the metal. It has made her allergic, prone to fits at any contact with ichorite, but also able to exert a strange control over the metal if she's willing to pay the price of spasms and hallucinations for hours afterwards. As Metal from Heaven opens, the workers have declared a strike. Her older sister is the spokesperson, demanding shorter hours, safer working conditions, and an investigation into the health of the lustertouched children. Chauncey's response is to send enforcer snipers to kill the workers, including the entirety of her family.
The girl sang, "Unalone toward dawn we go, toward the glory of the new morning." An enforcer shot her in the belly, and when she did not fall, her head.
Marney survives, fleeing into the city, swearing an impossible personal revenge against Yann Chauncey. An act of charity gets her a ticket on a train into the countryside. The woman who bought her ticket is a bandit who is on the train to rob it. Marney's ability to control ichorite allows her to help the bandits in return, winning her a place with the Highwayman's Choir who have been preying on the shipments of the rich and powerful and then disappearing into the hills. The Choir's secret is that the agoraphobic and paranoid Baron of the Fingerbluffs is dead and has been for years. He was killed by his staff, Hereafterist idealists, who have turned his remote territory into an anarchist commune and haven for pirates and bandits. This becomes Marney's home and the Choir becomes her family, but she never forgets her oath of revenge or the childhood friend she left behind in the piles of bodies and to whom this story is narrated. First, Clarke's writing is absolutely gorgeous.
We scaled the viny mountain jags at Montrose Barony's legal edge, the place where land was and wasn't Ignavia, Royston, and Drustland alike. There was a border but it was diffuse and hallucinatory, even more so than most. On legal papers and state maps there were harsh lines that squashed topography and sanded down the mountains into even hills in planter's rows, but here among the jutting rocks and craggy heather, the ground was lineless.
The rhythm of it, the grasp of contrast and metaphor, the word choice! That climactic word "lineless," with its echo of limitless. So good. Second, this is the rarest of books: a political fantasy that takes class and religion seriously and uses them for more than plot drivers. This is not at all our world, and the technology level is somewhat ambiguous, but the parallels to the Gilded Age and Progressive Era are unmistakable. The Hereafterists that Marney joins are political anarchists, not in the sense of alternative governance structures and political theory sanitized for middle-class liberals, but in the sense of Emma Goldman and Peter Kropotkin. The society they have built in the Fingerbluffs is temporary, threatened, and contingent, but it is sincere and wildly popular among the people who already lived there. Even beyond politics, class is a tangible force in this book. Marney is a factory worker and the child of factory workers. She barely knows how to read and doesn't magically learn over the course of the book. She has friends who are clever in the sense rewarded by politics and nobility, who navigate bureaucracies and political nuance, but that is not Marney's world. When, towards the end of the book, she has to deal with a gathering of high-class women, the contrast is stark, and she navigates that gathering only by being entirely unexpected. Perhaps the best illustration of the subtlety of this is the terminology in the book for lesbian. Marney is a crawly, which is a slur thrown at people like her (and one of the rare fictional slurs that work exactly as the author intended) but is also simply what she calls herself. Whether or not it functions as a slur depends on context, and the context is never hard to understand. The high-class lesbians she meets later are Lunarists, and react to crawly as a vile and insulting word. They use language to separate themselves from both the insult and from the social class that uses it. Language is an indication of culture and manners and therefore of morality, unlike deeds, which admit endless justifications.
Conversation was fleeting. Perdita managed with whomever stood near her, chipper about every prettiness she saw, the flitting butterflies, the dappled light between the leaves, the lushness and the fragrance of untamed land, and her walking companions took turns sharing in her delight. It was infectious, how happy she was. She was going to slaughter millions. She was going to skip like this all the while.
The handling of religion is perhaps even better. Marney was raised a Tullian, which sits alongside two other fleshed-out fictional religions and sketches of several more. Tullians tend to be conservative and patriarchal, and Marney has a realistically complicated relationship with faith: sticking with some Tullian worship practices and gestures because they're part of who she is, feeling a kinship to other Tullians, discarding beliefs that don't fit her, and revising others. Every major religion has a Hereafterist spin or reinterpretation that upends or reverses the parts of the religion that were used to prop up the existing social order and brings it more in line with Hereafterist ideals. We see the Tullian Hereafterist variation in detail, and as someone who has studied a lot of methods of reinterpreting Christianity, I was impressed by how well Clarke invents both a belief system and its revisionist rewrite. This is exactly how religions work in human history, but one almost never sees this subtlety in fantasy novels. Marney's allergy to ichorite causes her internal dialogue to dissolve into hallucinatory synesthesia when she's manipulating or exposed to it. Since that's most of the book, substantial portions read like drug trips with growing body horror. I normally hate this type of narration, so it's a sign of just how good Clarke's writing is that I tolerated it and even enjoyed parts. It helps that the descriptions are irreverent and often surprising, full of unexpected metaphors and sudden turns. It's very hard not to quote paragraph after paragraph of this book. Clarke is also doing a lot with gender that I don't feel qualified to comment in detail on, but it would not surprise me to see this book in the Otherwise Award recommendation list. I can think of three significant male characters, all of whom are well-done, but every other major character is female by at least some gender definition. Within that group, though, is huge gender diversity of the complicated and personal type that doesn't force people into defined boxes. Marney's sexuality is similarly unclassified and sometimes surprising. My one complaint is that I thought the sex scenes (which, to warn, are often graphic) fell into the literary fiction trap of being described so closely and physically that it didn't feel like anyone involved was actually enjoying themselves. (This is almost certainly a matter of personal taste.) I had absolutely no idea how Clarke was going to end this book, and the last couple of chapters caught me by surprise. I'm still not sure what I think about the climax. It's not the ending that I wanted, but one of the merits of this book is that it never did what I thought I wanted and yet made me enjoy the journey anyway. It is, at least, a genre ending, not a literary ending: The reader gets a full explanation of what is going on, and the setting is not static the way that it so often is in literary fiction. The characters can change the world, for good or for ill. The story felt frustrating and incomplete when I first finished it, but I haven't stopped thinking about this book and I think I like the shape of it a bit more now. It was certainly unexpected, at least by me. Clarke names Dhalgren as one of their influences in the acknowledgments, and yes, Metal from Heaven is that kind of book. This is the first 2024 novel I've read that felt like the kind of book that should be on award shortlists. I'm not sure it was entirely successful, and there are parts of it that I didn't like or that weren't for me, but it's trying to do something different and challenging and uncomfortable, and I think it mostly worked. And the writing is so good.
She looked like a mythic princess from the old woodcuts, who ruled nature by force of goodness and faith and had no legal power.
Metal from Heaven is not going to be everyone's taste. If you do not like literary fantasy, there is a real chance that you will hate this. I am very glad that I read it, and also am going to take a significant break from difficult books before I tackle another one. But then I'm probably going to try the Scapegracers series, because Clarke is an author I want to follow. Content notes: Explicit sex, including sadomasochistic sex. Political violence, mostly by authorities. Murdered children, some body horror, and a lot of serious injuries and death. Rating: 8 out of 10

kpcyrd: 2024 wrapped

Dear blog. This post is inspired by an old friend of mine who has been writing these for the past few years. I meant to do this for a while now, but ended up not preparing anything, so this post is me writing it from memory. There s likely stuff I forgot, me being gentle with myself I ll probably just permit myself to complete this list the next couple of days. I hate bragging, I try to not depend on external validation as much as possible, and being the anarcho-communist anti-capitalist that I am, I try to be content with knowing I m doing good in the background . I don t think people owe me for the work I did, I don t expect anything in return, and it s my way of giving back to the community and the people around me. Consider us even. That being said, I: Thanks to everybody who has been part of my human experience, past or present. Especially those who ve been closest. cheers,
kpcyrd

24 December 2024

Russ Allbery: Review: Number Go Up

Review: Number Go Up, by Zeke Faux
Publisher: Crown Currency
Copyright: 2023
Printing: 2024
ISBN: 0-593-44382-9
Format: Kindle
Pages: 373
Number Go Up is a cross between a history and a first-person account of investigative journalism around the cryptocurrency bubble and subsequent collapse in 2022. The edition I read has an afterward from June 2024 that brings the story up to date with Sam Bankman-Fried's trial and a few other events. Zeke Faux is a reporter for Bloomberg News and a fellow of New America. Last year, I read Michael Lewis's Going Infinite, a somewhat-sympathetic book-length profile of Sam Bankman-Fried that made a lot of people angry. One of the common refrains at the time was that people should read Number Go Up instead, and since I'm happy to read more about the absurdities of the cryptocurrency world, I finally got around to reading the other big crypto book of 2023. This is a good book, with some caveats that I am about to explain at absurd length. If you want a skeptical history of the cryptocurrency bubble, you should read it. People who think that it's somehow in competition with Michael Lewis's book or who think the two books disagree (including Faux himself) have profoundly missed the point of Going Infinite. I agree with Matt Levine: Both of these books are worth your time if this is the sort of thing you like reading about. But (much) more on Faux's disagreements with Lewis later. The frame of Number Go Up is Faux's quixotic quest to prove that Tether is a fraud. To review this book, I therefore need to briefly explain what Tether is. This is only the first of many extended digressions. One natural way to buy cryptocurrency would be to follow the same pattern as a stock brokerage account. You would deposit some amount of money into the account (or connect the brokerage account to your bank account), and then exchange money for cryptocurrency or vice versa, using bank transfers to put money in or take it out. However, there are several problems with this. One is that swapping cryptocurrency for money is awkward and sometimes expensive. Another is that holding people's investment money for them is usually highly regulated, partly for customer safety but also to prevent money laundering. These are often called KYC laws (Know Your Customer), and the regulation-hostile world of cryptocurrency didn't want to comply with them. Tether is a stablecoin, which means that the company behind Tether attempts to guarantee that one Tether is always worth exactly one US dollar. It is not a speculative investment like Bitcoin; it's a cryptocurrency substitute for dollars. People exchange dollars for Tether to get their money into the system and then settle all of their subsequent trades in Tether, only converting the Tether back to dollars when they want to take their money out of cryptocurrency entirely. In essence, Tether functions like the cash reserve in a brokerage account: Your Tether holdings are supposedly guaranteed to be equivalent to US dollars, you can withdraw them at any time, and because you can do so, you don't bother, instead leaving your money in the reserve account while you contemplate what new coin you want to buy. As with a bank, this system rests on the assurance that one can always exchange one Tether for one US dollar. The instant people stop believing this is true, people will scramble to get their money out of Tether, creating the equivalent of a bank run. Since Tether is not a regulated bank or broker and has no deposit insurance or strong legal protections, the primary defense against a run on Tether is Tether's promise that they hold enough liquid assets to be able to hand out dollars to everyone who wants to redeem Tether. (A secondary defense that I wish Faux had mentioned is that Tether limits redemptions to registered accounts redeeming more than $100,000, which is a tiny fraction of the people who hold Tether, but for most purposes this doesn't matter because that promise is sufficient to maintain the peg with the dollar.) Faux's firmly-held belief throughout this book is that Tether is lying. He believes they do not have enough money to redeem all existing Tether coins, and that rather than backing every coin with very safe liquid assets, they are using the dollars deposited in the system to make illiquid and risky investments. Faux never finds the evidence that he's looking for, which makes this narrative choice feel strange. His theory was tested when there was a run on Tether following the collapse of the Terra stablecoin. Tether passed without apparent difficulty, redeeming $16B or about 20% of the outstanding Tether coins. This doesn't mean Faux is wrong; being able to redeem 20% of the outstanding tokens is very different from being able to redeem 100%, and Tether has been fined for lying about its reserves. But Tether is clearly more stable than Faux thought it was, which makes the main narrative of the book weirdly unsatisfying. If he admitted he might be wrong, I would give him credit for showing his work even if it didn't lead where he expected, but instead he pivots to focusing on Tether's role in money laundering without acknowledging that his original theory took a serious blow. In Faux's pursuit of Tether, he wanders through most of the other elements of the cryptocurrency bubble, and that's the strength of this book. Rather than write Number Go Up as a traditional history, Faux chooses to closely follow his own thought processes and curiosity. This has the advantage of giving Faux an easy and natural narrative, something that non-fiction books of this type can struggle with, and it lets Faux show how confusing and off-putting the cryptocurrency world is to an outsider. The best parts of this book were the parts unrelated to Tether. Faux provides an excellent summary of the Axie Infinity speculative bubble and even traveled to the Philippines to interview people who were directly affected. He then wandered through the bizarre world of NFTs, and his first-hand account of purchasing one (specifically a Mutant Ape) to get entrance to a party (which sounded like a miserable experience I would pay money to get out of) really drives home how sketchy and weird cryptocurrency-related software and markets can be. He also went to El Salvador to talk to people directly about the country's supposed embrace of Bitcoin, and there's no substitute for that type of reporting to show how exaggerated and dishonest the claims of cryptocurrency adoption are. The disadvantage of this personal focus on Faux himself is that it sometimes feels tedious or sensationalized. I was much less interested in his unsuccessful attempts to interview the founder of Tether than Faux was, and while the digression into forced labor compounds in Cambodia devoted to pig butchering scams was informative (and horrific), I think Faux leaned too heavily on an indirect link to Tether. His argument is that cryptocurrency enables a type of money laundering that is particularly well-suited to supporting scams, but both scams and this type of economic slavery existed before cryptocurrency and will exist afterwards. He did not make a very strong case that Tether was uniquely valuable as a money laundering service, as opposed to a currently useful tool that would be replaced with some other tool should it go away. This part of the book is essentially an argument that money laundering is bad because it enables crime, and sure, to an extent I agree. But if you're going to put this much emphasis on the evils of money laundering, I think you need to at least acknowledge that many people outside the United States do not want to give US government, which is often openly hostile to them, veto power over their financial transactions. Faux does not. The other big complaint I have with this book, and with a lot of other reporting on cryptocurrency, is that Faux is sloppy with the term "Ponzi scheme." This is going to sound like nit-picking, but I think this sloppiness matters because it may obscure an ongoing a shift in cryptocurrency markets. A Ponzi scheme is not any speculative bubble. It is a very specific type of fraud in which investors are promised improbably high returns at very low risk and with safe principal. These returns are paid out, not via investment in some underlying enterprise, but by taking the money from new investments and paying it to earlier investors. Ponzi schemes are doomed because satisfying their promises requires a constantly increasing flow of new investors. Since the population of the world is finite, all Ponzi schemes are mathematically guaranteed to eventually fail, often in a sudden death spiral of ever-increasing promises to lure new investors when the investment stream starts to dry up. There are some Ponzi schemes in cryptocurrency, but most practices that are called Ponzi schemes are not. For example, Faux calls Axie Infinity a Ponzi scheme, but it was missing the critical elements of promised safe returns and fraudulently paying returns from the investments of later investors. It was simply a speculative bubble that people bought into on the assumption that its price would increase, and like any speculative bubble those who sold before the peak made money at the expense of those who bought at the peak. The reason why this matters is that Ponzi schemes are a self-correcting problem. One can decry the damage caused when they collapse, but one can also feel the reassuring certainty that they will inevitably collapse and prove the skeptics correct. The same is not true of speculative assets in general. You may think that the lack of an underlying economic justification for prices means that a speculative bubble is guaranteed to collapse eventually, but in the famous words of Gary Schilling, "markets can remain irrational a lot longer than you and I can remain solvent." One of the people Faux interviews explains this distinction to him directly:
Rong explained that in a true Ponzi scheme, the organizer would have to handle the "fraud money." Instead, he gave the sneakers away and then only took a small cut of each trade. "The users are trading between each other. They are not going through me, right?" Rong said. Essentially, he was arguing that by downloading the Stepn app and walking to earn tokens, crypto bros were Ponzi'ing themselves.
Faux is openly contemptuous of this response, but it is technically correct. Stepn is not a Ponzi scheme; it's a speculative bubble. There are no guaranteed returns being paid out of later investments and no promise that your principal is safe. People are buying in at price that you may consider irrational, but Stepn never promised you would get your money back, let alone make a profit, and therefore it doesn't have the exponential progression of a Ponzi scheme. One can argue that this is a distinction without a moral difference, and personally I would agree, but it matters immensely if one is trying to analyze the future of cryptocurrencies. Schemes as transparently unstable as Stepn (which gives you coins for exercise and then tries to claim those coins have value through some vigorous hand-waving) are nearly as certain as Ponzi schemes to eventually collapse. But it's also possible to create a stable business around allowing large numbers of people to regularly lose money to small numbers of sophisticated players who are collecting all of the winnings. It's called a poker room at a casino, and no one thinks poker rooms are Ponzi schemes or are doomed to collapse, even though nearly everyone who plays poker will lose money. This is the part of the story that I think Faux largely missed, and which Michael Lewis highlights in Going Infinite. FTX was a legitimate business that made money (a lot of money) off of trading fees, in much the same way that a casino makes money off of poker rooms. Lots of people want to bet on cryptocurrencies, similar to how lots of people want to play poker. Some of those people will win; most of those people will lose. The casino doesn't care. Its profit comes from taking a little bit of each pot, regardless of who wins. Bankman-Fried also speculated with customer funds, and therefore FTX collapsed, but there is no inherent reason why the core exchange business cannot be stable if people continue to want to speculate in cryptocurrencies. Perhaps people will get tired of this method of gambling, but poker has been going strong for 200 years. It's also important to note that although trading fees are the most obvious way to be a profitable cryptocurrency casino, they're not the only way. Wall Street firms specialize in finding creative ways to take a cut of every financial transaction, and many of those methods are more sophisticated than fees. They are so good at this that buying and selling stock through trading apps like Robinhood is free. The money to run the brokerage platform comes from companies that are delighted to pay for the opportunity to handle stock trades by day traders with a phone app. This is not, as some conspiracy theories would have you believe, due to some sort of fraudulent price manipulation. It is because the average person with a Robinhood phone app is sufficiently unsophisticated that companies that have invested in complex financial modeling will make a steady profit taking the other side of their trades, mostly because of the spread (the difference between offered buy and sell prices). Faux is so caught up in looking for Ponzi schemes and fraud that I think he misses this aspect of cryptocurrency's transformation. Wall Street trading firms aren't piling into cryptocurrency because they want to do securities fraud. They're entering this market because there seems to be persistent demand for this form of gambling, cryptocurrency markets reward complex financial engineering, and running a legal casino is a profitable business model. Michael Lewis appears as a character in this book, and Faux portrays him quite negatively. The root of this animosity appears to stem from a cryptocurrency conference in the Bahamas that Faux attended. Lewis interviewed Bankman-Fried on stage, and, from Faux's account, his questions were fawning and he praised cryptocurrencies in ways that Faux is certain he knew were untrue. From that point on, Faux treats Lewis as an apologist for the cryptocurrency industry and for Sam Bankman-Fried specifically. I think this is a legitimate criticism of Lewis's methods of getting close to the people he wants to write about, but I think Faux also makes the common mistake of assuming Lewis is a muckraking reporter like himself. This has never been what Lewis is interested in. He writes about people he finds interesting and that he thinks a reader will also find interesting. One can legitimately accuse him of being credulous, but that's partly because he's not even trying to do the same thing Faux is doing. He's not trying to judge; he's trying to understand. This shows when it comes to the parts of this book about Sam Bankman-Fried. Faux's default assumption is that everyone involved in cryptocurrency is knowingly doing fraud, and a lot of his research is looking for evidence to support the conclusion he had already reached. I don't think there's anything inherently wrong with that approach: Faux is largely, although not entirely, correct, and this type of hostile journalism is incredibly valuable for society at large. Upton Sinclair didn't start writing The Jungle with an open mind about the meat-packing industry. But where Faux and Lewis disagree on Bankman-Fried's motivations and intentions, I think Lewis has the much stronger argument. Faux's position is that Bankman-Fried always intended to steal people's money through fraud, perhaps to fund his effective altruism donations, and his protestations that he made mistakes and misplaced funds are obvious lies. This is an appealing narrative if one is looking for a simple villain, but Faux's evidence in support of this is weak. He mostly argues through stereotype: Bankman-Fried was a physics major and a Jane Street trader and therefore could not possibly be the type of person to misplace large amounts of money or miscalculate risk. If he wants to understand how that could be possible, he could read Going Infinite? I find it completely credible that someone with what appears to be uncontrolled, severe ADHD could be adept at trading and calculating probabilities and yet also misplace millions of dollars of assets because he wasn't thinking about them and therefore they stopped existing. Lewis made a lot of people angry by being somewhat sympathetic to someone few people wanted to be sympathetic towards, but Faux (and many others) are also misrepresenting his position. Lewis agrees that Bankman-Fried intentionally intermingled customer funds with his hedge fund and agrees that he lied about doing this. His only contention is that Bankman-Fried didn't do this to steal the money; instead, he invested customer money in risky bets that he thought would pay off. In support of this, Lewis made a prediction that was widely scoffed at, namely that much less of FTX's money was missing than was claimed, and that likely most or all of it would be found. And, well, Lewis was basically correct? The FTX bankruptcy is now expected to recover considerably more than the amount of money owed to creditors. Faux argues that this is only because the bankruptcy clawed back assets and cryptocurrencies have gone up considerably since the FTX bankruptcy, and therefore that the lost money was just replaced by unexpected windfall profits on other investments, but I don't think this point is as strong as he thinks it is. Bankman-Fried lost money on some of what he did with customer funds, made money on other things, and if he'd been able to freeze withdrawals for the year that the bankruptcy froze them, it does appear most of the money would have been recoverable. This does not make what he did legal or morally right, but no one is arguing that, only that he didn't intentionally steal money for his own personal gain or for effective altruism donations. And on that point, I don't think Faux is giving Lewis's argument enough credit. I have a lot of complaints about this book because I know way too much about this topic than anyone should probably know. I think Faux missed the plot in a couple of places, and I wish someone would write a book about where cryptocurrency markets are currently going. (Matt Levine's Money Stuff newsletter is quite good, but it's about all sorts of things other than cryptocurrency and isn't designed to tell a coherent story.) But if you know less about cryptocurrency and just want to hear the details of the run-up to the 2022 bubble, this is a great book for that. Faux is writing for people who are already skeptical and is not going to convince people who are cryptocurrency true believers, but that's fine. The details are largely correct (and extensively footnoted) and will satisfy most people's curiosity. Lewis's Going Infinite is a better book, though. It's not the same type of book at all, and it will not give you the broader overview of the cryptocurrency world. But if you're curious about what was going through the head of someone at the center of all of this chaos, I think Lewis's analysis is much stronger than Faux's. I'm happy I read both books. Rating: 8 out of 10

23 December 2024

Sahil Dhiman: Debian Mirrors Hierarchy

After finding AlmaLinux mirror sync capacity at Tier 0 (or Tier 1, however you look at it) is around 140 Gbps, I wanted to find source and hierarchy in Debian mirroring systems. There are two main types of mirrors in Debian - Debian package mirrors (for package installs and updates) and Debian CD mirrors (for ISO and others medias). Let s talk about package mirrors (and it s hierarchy) first.

Package mirror hierarchy Trace file was a good starting point for checking upstream for a package mirror in Debian. It resides at <URL>/debian/project/trace/_traces and shows flow of data. Sample trace file from jing.rocks s mirror. It showed, canonical source for packages is ftp-master.debian.org. Checking via https://db.debian.org/machines.cgi, showed it s fasolo.d.o hosted at Brown University, US. This serves as Master Archive Server , making it a Tier 0 mirror. It s entry mentions that it has 1 Gbps shared LAN connectivity (dated information?) but it only has to push to 3 other machines/sites. Side note - .d.o is .debian.org As shown on https://mirror-master.debian.org/status/mirror-hierarchy.html, the three sites are:
  • syncproxy2.eu.debian.org ie smit.d.o hosted by University of Twente, Netherlands with 2x10 Gbps connectivity.
  • syncproxy4.eu.debian.org ie schmelzer.d.o hosted by Conova in Austria with 2x10 Gbps connectivity.
  • syncproxy2.wna.debian.org - https://db.debian.org/machines.cgi entry mentions it being hosted at UBC here, but IP seems to be pointing to OSUOSL IP range as of now. IIRC few months ago, syncproxy2.wna.d.o was made to point to other host due to some issue (?). mirror-osuosl.d.o seems to be serving as syncproxy2.wna.d.o now. Bandwidth isn t explicitly mentioned but from my experience seeing bandwidths which other free software projects hosted at OSUOSL have, it would be atleast 10 Gbps and maybe more for Debian.

                     syncproxy2.eu.d.o (NL) ---> to the world
                    /
ftp-master.d.o (US) -- syncproxy4.eu.d.o (AT)  --> to the world 
                    \
                     syncproxy2.wna.d.o (US) --> to the world
A visualation of flow of package from ftp-master.d.o
These form the Debian Tier 1 mirror network, as all the mirrors sync from them. So Debian has atleast 50 Gbps+ capacity at Tier 1. A normal Debian user might never directly interact with any of these 3 machines, but every Debian package they run/download/install flows through these machines. Though, I m unsure what wna stands for in syncproxy2.wna.d.o. NA probably is North America and W is west (coast)? If you know, do let me know. After Tier 1, there are a few more syncproxies (detailed below). There are atleast 45 mirrors at Tier 2, updates for which are directly pushed from the three Tier 1 sync proxies. Most country mirrors i.e. ftp..debian.org are at Tier 2 too (barring a few like ftp.au.d.o, ftp.nz.do etc). Coming back to Sync proxies at Tier 2:
  • syncproxy3.wna.debian.org - gretchaninov.d.o which is marked as syncproxy2 on db.d.o (information dated). It s hosted in University of British Columbia, Canada, where a lot of Debian infrastructure including Salsa is hosted.
  • syncproxy.eu.debian.org - Croatian Academic and Research Network managed machine. CNAME directs to debian.carnet.hr.
  • syncproxy.au.debian.org - mirror-anu.d.o hosted by Australian National University with 100Mbps connectivity. Closest sync proxy for all Australian mirrors.
  • syncproxy4.wna.debian.org - syncproxy-aws-wna-01.d.o hosted in AWS, in US (according to GeoIP). IPv6 only (CNAME to syncproxy-aws-wna-01.debian.org. which only has an AAAA record, no A record). A m6g.2xlarge instance which has speeds upto 10 Gbps.
Coming back to https://mirror-master.debian.org/status/mirror-hierarchy.html, one can see chain extend till Tier 6 like in case of this mirror in AU which should add some latency for the updates from being pushed at ftp-master.d.o to them. Ideally, which shouldn t be a problem as https://www.debian.org/mirror/ftpmirror#when mentions The main archive gets updated four times a day . In my case, I get my updates from NITC mirror, so my updates flows from US > US > TW > IN > me in IN. CDNs have to internally manage cache purging too unlike normal mirrors which directly serve static file. Both deb.debian.org (sponsored by Fastly) and cdn-aws.deb.debian.org (sponsored by Amazon Cloudfront) sync from following CDN backends: See deb.d.o trace file and cdn-aws.deb.d.o trace file. (Thanks to Philipp Kern for the heads up here.)

CD image mirrors Hierarchy Till now, I have only talked about Debian package mirrors. When you see /debian directory on various mirrors, they re usually for package install and updates. If you want to grab the latest (and greatest) Debian ISO, you go to Debian CD (as they re still called) mirror site. casulana.d.o is mentioned as CD builder site hosted by Bytemark while pettersson-ng.d.o is mentioned as CD publishing server hosted at Academic Computer Club in Ume , Sweden. Primary download site for Debian CD when you click download on debian.org homepage is https://cdimage.debian.org/debian-cd/ is hosted here as well. This essentially becomes Tier 0 mirror for Debian CD. All Debian CD mirrors are downstream to it.
pettersson-ng.d.o / cdimage.d.o (SE) ---> to the world
A visualation of flow of Debian CD from cdimage.d.o
Academic Computer Club s mirror setup uses a combination of multiple machines (called frontends and offloading servers) to load balance requests. Their document setup is a highly recommended read. Also, in that document, they mention , All machines are reachable via both IPv4 and IPv6 and connected with 10 or 25 gigabit Ethernet, external bandwidth available is 200 gigabit/s. For completeness sake, following mirror (or mirror systems) exists too for Debian: Debian heavily rely on various organizations to donate resources (hosting and hardware) to distribute and update Debian. Compiling above information made me thankful to all these organizations. Many thanks to DSA and mirror team as well for managing these stuffs. I relied heavily on https://db.debian.org/machines.cgi which seems to be manually updated, so things might have changed along the way. If anything looks amiss, feel free to ping.

20 December 2024

Michael Prokop: Grml 2024.12 codename Adventgrenze

Picture with metrics of three user profiles on GitHub.com, with many contributions especially in the last quarter of the year We did it again ! Just in time, we re excited to announce the release of Grml stable version 2024.12, code-named Adventgrenze ! (If you re not familiar with Grml, it s a Debian-based live system tailored for system administrators.) This new release is built on Debian trixie, and for the first time, we re introducing support for 64-bit ARM CPUs (arm64 architecture)! I m incredibly proud of the hard work that went into this release. A significant amount of behind-the-scenes effort went into reworking our infrastructure and redesigning the build process. Special thanks to Chris and Darsha our Grml developer days in November and December were a blast! For a detailed overview of the changes between releases 2024.02 and 2024.12, check out our official release announcement. And, as always, after a release comes the next one exciting improvements are already in the works! BTW: recently we also celebrated 20(!) years of Grml Releases. If you re a Grml and or grml-zsh user, please join us in celebrating and send us a postcard!

17 December 2024

Gunnar Wolf: The science of detecting LLM-generated text

This post is a review for Computing Reviews for The science of detecting LLM-generated text , a article published in Communications of the ACM
While artificial intelligence (AI) applications for natural language processing (NLP) are no longer something new or unexpected, nobody can deny the revolution and hype that started, in late 2022, with the announcement of the first public version of ChatGPT. By then, synthetic translation was well established and regularly used, many chatbots had started attending users requests on different websites, voice recognition personal assistants such as Alexa and Siri had been widely deployed, and complaints of news sites filling their space with AI-generated articles were already commonplace. However, the ease of prompting ChatGPT or other large language models (LLMs) and getting extensive answers its text generation quality is so high that it is often hard to discern whether a given text was written by an LLM or by a human has sparked significant concern in many different fields. This article was written to present and compare the current approaches to detecting human- or LLM-authorship in texts. The article presents several different ways LLM-generated text can be detected. The first, and main, taxonomy followed by the authors is whether the detection can be done aided by the LLM s own functions ( white-box detection ) or only by evaluating the generated text via a public application programming interface (API) ( black-box detection ). For black-box detection, the authors suggest training a classifier to discern the origin of a given text. Although this works at first, this task is doomed from its onset to be highly vulnerable to new LLMs generating text that will not follow the same patterns, and thus will probably evade recognition. The authors report that human evaluators find human-authored text to be more emotional and less objective, and use grammar to indicate the tone of the sentiment that should be used when reading the text a trait that has not been picked up by LLMs yet. Human-authored text also tends to have higher sentence-level coherence, with less term repetition in a given paragraph. The frequency distribution for more and less common words is much more homogeneous in LLM-generated texts than in human-written ones. White-box detection includes strategies whereby the LLMs will cooperate in identifying themselves in ways that are not obvious to the casual reader. This can include watermarking, be it rule based or neural based; in this case, both processes become a case of steganography, as the involvement of a LLM is explicitly hidden and spread through the full generated text, aiming at having a low detectability and high recoverability even when parts of the text are edited. The article closes by listing the authors concerns about all of the above-mentioned technologies. Detecting an LLM, be it with or without the collaboration of the LLM s designers, is more of an art than a science, and methods deemed as robust today will not last forever. We also cannot assume that LLMs will continue to be dominated by the same core players; LLM technology has been deeply studied, and good LLM engines are available as free/open-source software, so users needing to do so can readily modify their behavior. This article presents itself as merely a survey of methods available today, while also acknowledging the rapid progress in the field. It is timely and interesting, and easy to follow for the informed reader coming from a different subfield.

16 December 2024

Russ Allbery: Review: Finders

Review: Finders, by Melissa Scott
Series: Firstborn, Lastborn #1
Publisher: Candlemark & Gleam
Copyright: 2018
ISBN: 1-936460-87-4
Format: Kindle
Pages: 409
Finders is a far future science fiction novel with cyberpunk vibes. It is the first of a series, but the second (and, so far, only other) book of the series is a prequel. It stands alone reasonably well (more on that later). Cassilde Sam is a salvor. That means she specializes in exploring ancient wrecks and ruins left behind by the Ancients and salvaging materials that can be reused. The most important of those are what are called Ancestral elements: BLUE, which can hold programming; GOLD, which which reacts to BLUE instructions; RED, which produces actions or output; and GREEN, the rarest and most valuable, which powers everything else. Cassilde and her partner Dai Winter file claims on newly-discovered or incompletely salvaged Ancestor sites and then extract elemental material and anything else of value in their small salvage ship. Cassilde is also dying. She has Lightman's, an incurable degenerative disease that can only be treated with ever-increasing quantities of GREEN. It's hard to sleep, hard to get warm, hard to breathe, and eventually she'll run out of money to pay for the GREEN and she'll die. To push that day off into the future, she and Dai need work. The good news is that the wreckage of a new Ancestor sky palace was discovered in a long orbit and will create enough salvage work for every experienced salvor in the system. The bad news is that they're not qualified to bid on it. They need a scholar with a class-one license to bid on the best sections, and they haven't had a reliable scholar since their former partner and lover Summerland Ashe picked the opposite side in the Troubles and left the Fringe for the Entente, the more densely settled and connected portion of human space. But, unexpectedly and suspiciously, Ashe may be back and offering to work with them again. So, first, I love this setting. This is far from the first SF novel that is set in the aftermath of a general collapse of human civilization and revolving around discovering lost mysteries. Most examples of that genre are post-apocalyptic novels limited to Earth or the local solar system, but Kate Elliott's Unconquerable Sun comes immediately to mind. It's also not the first space archaeology series I've read; Kristine Kathyrn Rusch's story series starting with "Diving into the Wreck" also came to mind. But I don't recall the last time I've seen the author sell the setting so effectively. This is a world with starships and spaceports and clearly advanced technology, but it feels like a post-collapse society that's built on ruins. It's not just that technology runs on half-understood Ancestral elements and states fight over control of debris fields. It's also that the society repurposes Ancestral remnants in ways that both they and the reader know weren't originally intended, and that sometimes are more ingenious or efficient than how the Ancestors probably used them. There's a creative grittiness here that reminds me of good cyberpunk. It's not just good atmospheric writing, though. Scott makes a world-building decision that is going to sound trivial when I say it, but that has brilliant implications for the rest of the setting. There was not just one collapse; there were two. The Ancestor civilization, presumed to be the first human civilization, has passed into myth, quite literally when it comes to the stories around its downfall in the aftermath of a war against AIs. After the Ancestors came the Successors, who followed a similar salvage and rebuild approach and got as far as inventing their own warp drive technology that was based on but different than the Ancestor technology. Then they also collapsed, leaving their adapted technology and salvage operations layered over Ancestor sites. Cassilde's civilization is the third human starfaring civilization, and it is very specifically the third, neither the second nor one of dozens. This has so many small but effective implications that improve this story. A fall happened twice, so it feels like a pattern that makes Cassilde's civilization paranoid, but it happened for two very different reasons, so there is room to argue against it being a pattern. Salvage is harder because of the layering of Ancestor and Successor activity. Successors had their own way of controlling technology that is not accessible to Cassilde and her crew but is also not how the technology was intended to be used, which sends small ripples of interesting complexity through the background. And salvors are competing not only against each other but also against Successor salvage operations for which they have fragmentary records. It's a beautifully effective touch. Melissa Scott has been publishing science fiction for forty years, and it shows in this book. The protagonists are older characters: established professionals with resource problems but also social connections and an earned reputation, people who are trying to do a job and live their lives, not change the world. The writing is competent, deft, and atmospheric, with the confidence of long practice, but it also has the feel of an earlier era of science fiction. I mentioned the cyberpunk influence, which shows in the grittiness of the descriptions, the marginality of the characters in society, and the background theme of repurposing and reusing technology in unintended ways. This is the sort of book that feels solidly in the center of science fiction, without the genre mixing into either fantasy or romance that has become somewhat more common, and also without the dramatics of space opera (although the reader discovers that the stakes of this novel may be higher than anyone realized). And yet, so much of this book is about navigating a complicated romantic relationship, and that's where the story structure felt a bit odd. Cassilde, Dai, and Ashe were a polyamorous triad (polyamory also shows up in Scott's excellent Roads of Heaven series), and much of the first third of the book deals with the fracturing of trust with Ashe and their renegotiation of that relationship given his return. This is refreshingly written as the thoughtful interaction of three adults who take issues of trust seriously, but that also means it's much less dramatic than it sounds, and that means this book starts exceptionally slow. Scott is going somewhere, and the slow build became engrossing around the midpoint of the book, but I had to fight to stick with it at the start. About 80% of the way through this book, I had no idea how Scott was going to wrap things up in the pages remaining and was bracing myself for some sort of series cliffhanger. This is not what happens; the plot is not fully resolved in every detail, but it reaches a conclusion of sorts that does not mandate a sequel. I did think the end was a little bit unsatisfying, though, and I want another book that explores the implications of the ending. I think it would have to be a much different book, and the tonal shift might be stark. I've had this book on my to-read list for a while and kept putting it off because I wasn't sure I was in the mood for something precarious and gritty. This turned out to be an accurate worry: this is literally a book about salvaging the pieces of something full of wonders inextricably connected to dangers. You have to be in a cyberpunk sort of mood. But I've never read a bad Melissa Scott book, and this is no exception. The simplicity and ALL-CAPSNESS of the Ancestral elements grated a bit, but apart from that, the world-building is exceptional and well worth the trip. Recommended, although be warned that, if you're like me, it may not grab you from the first page. Followed by Fallen, but that book is a prequel that does not share any protagonists. Content notes: disability and degenerative illness in a universe where magical cures are possible, so be warned if that specific thematic combination is not what you're looking for. Rating: 7 out of 10

13 December 2024

Emanuele Rocca: Murder Mystery: GCC Builds Failing After sbuild Refactoring

This is the story of an investigation conducted by Jochen Sprickerhof, Helmut Grohne, and myself. It was true teamwork, and we would have not reached the bottom of the issue working individually. We think you will find it as interesting and fun as we did, so here is a brief writeup. A few of the steps mentioned here took several days, others just a few minutes. What is described as a natural progression of events did not always look very obvious at the moment at all.
Let us go through the Six Stages of Debugging together.

Stage 1: That cannot happen
Official Debian GCC builds start failing on multiple architectures in late November.
The build error happens on the build servers when running the testuite, but we know this cannot happen. GCC builds are not meant to fail in case of testsuite failures! Return codes are not making the build fail, make is being called with -k, it just cannot happen.
A lot of the GCC tests are always failing in fact, and an extensive log of the results is posted to the debian-gcc mailing list, but the packages always build fine regardless.
On the build daemons, build failures take several hours.

Stage 2: That does not happen on my machine
Building on my machine running Bookworm is just fine. The Build Daemons run Bookworm and use a Sid chroot for the build environment, just like I am. Same kernel.
The only obvious difference between my setup and the Debian buildds is that I am using sbuild 0.85.0 from bookworm, and the buildds have 0.86.3~bpo12+1 from bookworm-backports. Trying again with 0.86.3~bpo12+1, the build fails on my system too. The build daemons were updated to the bookworm-backports version of sbuild at some point in late November. Ha.

Stage 3: That should not happen
There are quite a few sbuild versions in between 0.85.0 and 0.86.3~bpo12+1, but looking at recent sbuild bugs shows that sbuild 0.86.0 was breaking "quite a number of packages". Indeed, with 0.86.0 the build still fails. Trying the version immediately before, 0.85.11, the build finishes correctly. This took more time than it sounds, one run including the tests takes several hours. We need a way to shorten this somehow.
The Debian packaging of GCC allows to specify which languages you may want to skip, and by default it builds Ada, Go, C, C++, D, Fortran, Objective C, Objective C++, M2, and Rust. When running the tests sequentially, the build logs stop roughly around the tests of a runtime library for D, libphobos. So can we still reproduce the failure by skipping everything except for D? With DEB_BUILD_OPTIONS=nolang=ada,go,c,c++,fortran,objc,obj-c++,m2,rust the build still fails, and it fails faster than before. Several minutes, not hours. This is progress, and time to file a bug. The report contains massive spoilers, so no link. :-)

Stage 4: Why does that happen?
Something is causing the build to end prematurely. It s not the OOM killer, and the kernel does not have anything useful to say in the logs. Can it be that the D language tests are sending signals to some process, and that is what s killing make ? We start tracing signals sent with bpftrace by writing the following script, signals.bt:
tracepoint:signal:signal_generate  
    printf("%s PID %d (%s) sent signal %d to PID %d\n", comm, pid, args->sig, args->pid);
 
And executing it with sudo bpftrace signals.bt.
The build takes its sweet time, and it fails. Looking at the trace output there s a suspicious process.exe terminating stuff.
process.exe (PID: 2868133) sent signal 15 to PID 711826
That looks interesting, but we have no clue what PID 711826 may be. Let s change the script a bit, and trace signals received as well.
tracepoint:signal:signal_generate  
    printf("PID %d (%s) sent signal %d to %d\n", pid, comm, args->sig, args->pid);
 
tracepoint:signal:signal_deliver  
    printf("PID %d (%s) received signal %d\n", pid, comm, args->sig);
 
The working version of sbuild was using dumb-init, whereas the new one features a little init in perl. We patch the current version of sbuild by making it use dumb-init instead, and trace two builds: one with the perl init, one with dumb-init.
Here are the signals observed when building with dumb-init.
PID 3590011 (process.exe) sent signal 2 to 3590014
PID 3590014 (sleep) received signal 9
PID 3590011 (process.exe) sent signal 15 to 3590063
PID 3590063 (std.process tem) received signal 9
PID 3590011 (process.exe) sent signal 9 to 3590065
PID 3590065 (std.process tem) received signal 9
And this is what happens with the new init in perl:
PID 3589274 (process.exe) sent signal 2 to 3589291
PID 3589291 (sleep) received signal 9
PID 3589274 (process.exe) sent signal 15 to 3589338
PID 3589338 (std.process tem) received signal 9
PID 3589274 (process.exe) sent signal 9 to 3589340
PID 3589340 (std.process tem) received signal 9
PID 3589274 (process.exe) sent signal 15 to 3589341
PID 3589274 (process.exe) sent signal 15 to 3589323
PID 3589274 (process.exe) sent signal 15 to 3589320
PID 3589274 (process.exe) sent signal 15 to 3589274
PID 3589274 (process.exe) received signal 9
PID 3589341 (sleep) received signal 9
PID 3589273 (sbuild-usernsex) sent signal 9 to 3589320
PID 3589273 (sbuild-usernsex) sent signal 9 to 3589323
There are a few additional SIGTERM being sent when using the perl init, that s helpful. At this point we are fairly convinced that process.exe is worth additional inspection. The source code of process.d shows something interesting:
1221 @system unittest
1222  
[...]
1247     auto pid = spawnProcess(["sleep", "10000"],
[...]
1260     // kill the spawned process with SIGINT
1261     // and send its return code
1262     spawn((shared Pid pid)  
1263         auto p = cast() pid;
1264         kill(p, SIGINT);
So yes, there s our sleep and the SIGINT (signal 2) right in the unit tests of process.d, just like we have observed in the bpftrace output.
Can we study the behavior of process.exe in isolation, separatedly from the build? Indeed we can. Let s take the executable from a failed build, and try running it under /usr/libexec/sbuild-usernsexec.
First, we prepare a chroot inside a suitable user namespace:
unshare --map-auto --setuid 0 --setgid 0 mkdir /tmp/rootfs
cd /tmp/rootfs
cat /home/ema/.cache/sbuild/unstable-arm64.tar   unshare --map-auto --setuid 0 --setgid 0 tar xf  -
unshare --map-auto --setuid 0 --setgid 0 mkdir /tmp/rootfs/whatever
unshare --map-auto --setuid 0 --setgid 0 cp process.exe /tmp/rootfs/
Now we can run process.exe on its own using the perl init, and trace signals at will:
/usr/libexec/sbuild-usernsexec --pivotroot --nonet u:0:100000:65536  g:0:100000:65536 /tmp/rootfs ema /whatever -- /process.exe
We can compare the behavior of the perl init vis-a-vis the one using dumb-init in milliseconds instead of minutes.

Stage 5: Oh, I see.
Why does process.exe send more SIGTERMs when using the perl init is now the big question. We have a simple reproducer, so this is where using strace becomes possible.
sudo strace --user ema --follow-forks -o sbuild-dumb-init.strace ./sbuild-usernsexec-dumb-init --pivotroot --nonet u:0:100000:65536  g:0:100000:65536 /tmp/dumbroot ema /whatever -- /process.exe
We start comparing the strace output of dumb-init with that of perl-init, looking in particular for different calls to kill.
Here is what process.exe does under dumb-init:
3593883 kill(-2, SIGTERM)               = -1 ESRCH (No such process)
No such process. Under perl-init instead:
3593777 kill(-2, SIGTERM <unfinished ...>
The process is there under perl-init!
That is a kill with negative pid. From the kill(2) man page:
If pid is less than -1, then sig is sent to every process in the process group whose ID is -pid.
It would have been very useful to see this kill with negative pid in the output of bpftrace, why didn t we? The tracepoint used, tracepoint:signal:signal_generate, shows when signals are actually being sent, and not the syscall being called. To confirm, one can trace tracepoint:syscalls:sys_enter_kill and see the negative PIDs, for example:
PID 312719 (bash) sent signal 2 to -312728
The obvious question at this point is: why is there no process group 2 when using dumb-init?

Stage 6: How did that ever work?
We know that process.exe sends a SIGTERM to every process in the process group with ID 2. To find out what this process group may be, we spawn a shell with dumb-init and observe under /proc PIDs 1, 16, and 17. With perl-init we have 1, 2, and 17. When running dumb-init, there are a few forks before launching the program, explaining the difference. Looking at /proc/2/cmdline we see that it s bash, ie. the program we are running under perl-init. When building a package, that is dpkg-buildpackage itself.
The test is accidentally killing its own process group.
Now where does this -2 come from in the test?
2363     // Special values for _processID.
2364     enum invalid = -1, terminated = -2;
Oh. -2 is used as a special value for PID, meaning "terminated". And there s a call to kill() later on:
2694     do   s = tryWait(pid);   while (!s.terminated);
[...]
2697     assertThrown!ProcessException(kill(pid));
What sets pid to terminated you ask?
Here is tryWait:
2568 auto tryWait(Pid pid) @safe
2569  
2570     import std.typecons : Tuple;
2571     assert(pid !is null, "Called tryWait on a null Pid.");
2572     auto code = pid.performWait(false);
And performWait:
2306         _processID = terminated;
The solution, dear reader, is not to kill.
PS: the bug report with spoilers for those interested is #1089007.

7 December 2024

Russ Allbery: Review: Dark Minds

Review: Dark Minds, by Michelle Diener
Series: Class 5 #3
Publisher: Eclipse
Copyright: 2016
ISBN: 0-6454658-5-2
Format: Kindle
Pages: 328
Dark Minds is the third book in the self-published Class 5 science-fiction romance series. It is a sequel to Dark Deeds and should not be read out of order, but as with the other books of the series, it follows new protagonists. Imogen, like another Earth woman before her, was kidnapped by a Class 5 for experimentation by the Tecran. She was subsequently transferred to a research facility and has been held in cages and secret military bases for over three months. As the story opens, she had been imprisoned in a Tecran transport for a couple of weeks when the transport is taken by the Krik who kill all of the Tecran crew. Next stop: another Class 5. That same Class 5 is where the Krik are taking Camlar Kalor and his team. Cam is a Grih investigator from the United Council, sent to look into reports of a second Earth woman rescued from a Garmman ship. (The series reader will recognize this as the plot of Dark Deeds.) Now he's trapped with the crews of a bunch of other random ships in the cargo hold of a Class 5 with Krik instead of Tecran roaming the halls, apparently at the behest of the ship's banned AI, and a mysterious third Earth woman already appears to be befriending it. Imogen is the woman Fiona saw signs of in Dark Deeds. She's had a rougher time than the protagonists of the previous two books in this series, and she's been dropped into a less stable situation. The Class 5 she's brought to at the start of the story is far more suspicious (with quite a lot of cause) and somewhat more hostile than the AIs we've encountered previously. The rest of the story formula is roughly the same, though: hunky Grih officer with a personality completely indistinguishable from the hunky Grih officers in the previous two books, an AI with a sketchy concept of morality that desperately needs an ally, the Grih obsession with singing, the eventual discovery of useful armor and weaponry that Imogen can use to surprise people, and more political maneuvering over the Sentient Beings Agreement. This entry in the series mostly abandons the Grih shock and horror at how badly the Earth women have been treated. This makes sense, given how dangerous Earth women have proven over the course of this series, and I like that Diener is changing the political dynamics as the story develops. I do sometimes miss that appalled anger of Dark Horse, but Dark Minds focuses more on the politics and corridor fighting of a tense multi-sided stand-off. I found the action more gripping in Dark Minds than in Dark Deeds, and I liked Imogen more as a character than Fiona. She doesn't have the delightfully calm competence of Rose from Dark Horse, but she's a bit more hardened, a bit more canny, and is better at taking control of situations. I also like that Diener avoids simplistically pairing Earth women off with Class 5s. The series plot is progressing faster than I had expected, and that gives this book a somewhat different shape than the previous ones. Cam is probably the least interesting of the men in this series so far and appears to exist solely to take up the man-shaped hole in the plot. This is not a great series for gender roles; thankfully, the romance is a small part of the plot and largely ignorable. The story is about the women and the AIs, and all of the women and most of the AIs of the previous books make an appearance. It's clear they're forming an alliance whether the Grih like it or not, and that part of the story was very satisfying. Up to this book, this series had been all feel-good happy endings. I will risk the small spoiler and warn that this is not true to the same degree here, so you may not want to read this one if you want something entirely fluffy, light, and positive (inasmuch as a series involving off-screen experimentation on humans can be fluffy, light, and positive). That caught me by surprise in a way I didn't entirely like, and I wish Diener had stuck with the entirely positive tone. Other than that, though, this was fun, light, readable entertainment. It's not going to win any literary awards, it's formulaic, the male protagonist comes from central casting, and the emphasis by paragraph break is still a bit grating in places, but I will probably pick up the next book when I'm in the mood for something light. Dark Minds is an improvement over book two, which bodes well for the rest of the series. Followed by Dark Matters. Rating: 7 out of 10

6 December 2024

Dirk Eddelbuettel: RcppArmadillo 14.2.2-1 on CRAN: Small Upstream Fixes

armadillo image Armadillo is a powerful and expressive C++ template library for linear algebra and scientific computing. It aims towards a good balance between speed and ease of use, has a syntax deliberately close to Matlab, and is useful for algorithm development directly in C++, or quick conversion of research code into production environments. RcppArmadillo integrates this library with the R environment and language and is widely used by (currently) 1197 other packages on CRAN, downloaded 37.5 million times (per the partial logs from the cloud mirrors of CRAN), and the CSDA paper (preprint / vignette) by Conrad and myself has been cited 605 times according to Google Scholar. Conrad released a minor version 14.2.2 yesterday. This followed a bit of recent work a few of us did in the ensmallen and mlpack repositories following the [14.2.0 release]. Use of (member functions) .min(index) and .max(index) was deprecated in Armadillo in favor of .index_mix() and .index_max(). By now ensmallen and mlpack have been updated at CRAN. To add some spice, CRAN emailed that the (very much unreleased as of now, but coming likely next spring) gcc-15 was unhappy with RcppArmadillo due to some Armadillo code. This likely related to the listed gcc-15 C++ change about Qualified name lookup failure into the current instantiation . Anyway, Conrad fixed it within days and that change too is part of this new version (as is a small behaviour normalization between the two indexing methods that matters in case of ties, this was in 14.2.1). The changes since the last CRAN release are summarised below.

Changes in RcppArmadillo version 14.2.2-1 (2024-12-05)
  • Upgraded to Armadillo release 14.2.2 (Smooth Caffeine)
    • Workarounds for regressions in pre-release versions of GCC 15
    • More selective detection of symmetric/hermitian matrices by various functions

Changes in RcppArmadillo version 14.2.1-1 (2024-11-24) (GitHub Only)
  • Upgraded to Armadillo release 14.2.1 (Smooth Caffeine)
    • Fix for index_min() and index_max() to ensure that the first index of equal extremum values is found

Courtesy of my CRANberries, there is a diffstat report relative to previous release. More detailed information is on the RcppArmadillo page. Questions, comments etc should go to the rcpp-devel mailing list off the Rcpp R-Forge page. If you like this or other open-source work I do, you can sponsor me at GitHub.

This post by Dirk Eddelbuettel originated on his Thinking inside the box blog. Please report excessive re-aggregation in third-party for-profit settings.

2 December 2024

Bits from Debian: Bits from the DPL

This is bits from DPL for November. MiniDebConf Toulouse I had the pleasure of attending the MiniDebConf in Toulouse, which featured a range of engaging talks, complementing those from the recent MiniDebConf in Cambridge. Both events were preceded by a DebCamp, which provided a valuable opportunity for focused work and collaboration. DebCamp During these events, I participated in numerous technical discussions on topics such as maintaining long-neglected packages, team-based maintenance, FTP master policies, Debusine, and strategies for separating maintainer script dependencies from runtime dependencies, among others. I was also fortunate that members of the Publicity Team attended the MiniDebCamp, giving us the opportunity to meet in person and collaborate face-to-face. Independent of the ongoing lengthy discussion on the Debian Devel mailing list, I encountered the perspective that unifying Git workflows might be more critical than ensuring all packages are managed in Git. While I'm uncertain whether these two questions--adopting Git as a universal development tool and agreeing on a common workflow for its use--can be fully separated, I believe it's worth raising this topic for further consideration. Attracting newcomers In my own talk, I regret not leaving enough time for questions--my apologies for this. However, I want to revisit the sole question raised, which essentially asked: Is the documentation for newcomers sufficient to attract new contributors? My immediate response was that this question is best directed to new contributors themselves, as they are in the best position to identify gaps and suggest improvements that could make the documentation more helpful. That said, I'm personally convinced that our challenges extend beyond just documentation. I don't get the impression that newcomers are lining up to join Debian only to be deterred by inadequate documentation. The issue might be more about fostering interest and engagement in the first place. My personal impression is that we sometimes fail to convey that Debian is not just a product to download for free but also a technical challenge that warmly invites participation. Everyone who respects our Code of Conduct will find that Debian is a highly diverse community, where joining the project offers not only opportunities for technical contributions but also meaningful social interactions that can make the effort and time truly rewarding. In several of my previous talks (you can find them on my talks page just search for "team," and don't be deterred if you see "Debian Med" in the title; it's simply an example), I emphasized that the interaction between a mentor and a mentee often plays a far more significant role than the documentation the mentee has to read. The key to success has always been finding a way to spark the mentee's interest in a specific topic that resonates with their own passions. Bug of the Day In my presentation, I provided a brief overview of the Bug of the Day initiative, which was launched with the aim of demonstrating how to fix bugs as an entry point for learning about packaging. While the current level of interest from newcomers seems limited, the initiative has brought several additional benefits. I must admit that I'm learning quite a bit about Debian myself. I often compare it to exploring a house's cellar with a flashlight you uncover everything from hidden marvels to things you might prefer to discard. I've also come across traces of incredibly diligent people who have invested their spare time polishing these hidden treasures (what we call NMUs). The janitor, a service in Salsa that automatically updates packages, fits perfectly into this cellar metaphor, symbolizing the ongoing care and maintenance that keep everything in order. I hadn't realized the immense amount of silent work being done behind the scenes--thank you all so much for your invaluable QA efforts. Reproducible builds It might be unfair to single out a specific talk from Toulouse, but I'd like to highlight the one on reproducible builds. Beyond its technical focus, the talk also addressed the recent loss of Lunar, whom we mourn deeply. It served as a tribute to Lunar's contributions and legacy. Personally, I've encountered packages maintained by Lunar and bugs he had filed. I believe that taking over his packages and addressing the bugs he reported is a meaningful way to honor his memory and acknowledge the value of his work. Advent calendar bug squashing I d like to promote an idea originally introduced by Thorsten Alteholz, who in 2011 proposed a Bug Squashing Advent Calendar for the Debian Med team. (For those unfamiliar with the concept of an Advent Calendar, you can find an explanation on Wikipedia.) While the original version included a fun graphical element which we ve had to set aside due to time constraints (volunteers, anyone?) we ve kept the tradition alive by tackling one bug per day from December 1st to 24th each year. This initiative helps clean up issues that have accumulated over the year. Regardless of whether you celebrate the concept of Advent, I warmly recommend this approach as a form of continuous bug-squashing party for every team. Not only does it contribute to the release readiness of your team s packages, but it s also an enjoyable and bonding activity for team members. Best wishes for a cheerful and productive December
Andreas.

Next.