| 27 Apr 2026 |
lillecarl | * ┃ # Submit build to pynixd
┃
┃ $ timeout 60 nix build --file ~/Code/pynixd/tests/nix simple --store /tmp/pynixd-stores/pynixkube --max-jobs 0 --builders "ssh-ng://nix@204.168.233.183:2222 x86_64-linux - 10" 2>&1
┃
┃ this derivation will be built:
┃ /nix/store/lsbmv7xvx6qrzqvc7lim5gj5kzgxg5vk-test-1777325199.drv
┃ these 50 paths will be fetched (21.7 MiB download, 104.4 MiB unpacked):
┃ /nix/store/742ik41xpll83mwfvmrlcv5h0cfd929n-ShellCheck-0.11.0
┃ /nix/store/4sabfgpxkbv6w3mvk0wil50vdi37m9r8-acl-2.3.2
┃ /nix/store/p2yk6q3bhcz1d0wlmk907ysj4l95ak7y-attr-2.5.2
┃ /nix/store/v8sa6r6q037ihghxfbwzjj4p59v2x0pv-bash-5.3p9
┃ /nix/store/91nr5z1cwr1qch2wl58dppl02zrnlza3-brotli-1.2.0-lib
┃ /nix/store/vi21p90v3kj509rx00r101xcqa6za76b-bzip2-1.0.8
┃ /nix/store/by6npbc13ly5a1zgsqghv7wv5bklyiny-bzip2-1.0.8-bin
┃ …
┃
┃ Click to expand
┃
Build succeeded end-to-end. Let me verify the full flow in the pynixd logs:
┃
┃ # Check full build flow in logs
┃
┃ $ stern -n nixkube pynixd-0 --no-follow --color never 2>/dev/null | grep -E "dynamic_features|builder_registered|builder_job_created|build_completed|adding_store_dynamically" | tail -10
┃
┃ pynixd-0 pynixd {"features": {"x86_64-linux": "{'nixos-test', 'big-parallel', 'benchmark'}"}, "event": "dynamic_features_registered", "timestamp": "2026-04-27T21:25:31.774396Z"}
┃ pynixd-0 pynixd {"job": "nixkube-builder-aafebb9f", "event": "builder_job_created", "timestamp": "2026-04-27T21:26:46.929259Z"}
┃ pynixd-0 pynixd {"store_id": "builder-nixkube-builder-aafebb9f-hk6zq", "dynamic": true, "event": "adding_store_dynamically", "timestamp": "2026-04-27T21:27:22.888235Z"}
┃ pynixd-0 pynixd {"store_id": "builder-nixkube-builder-aafebb9f-hk6zq", "pod_ip": "10.133.1.108", "event": "builder_registered", "timestamp": "2026-04-27T21:27:22.888318Z"}
┃ pynixd-0 pynixd {"build_id": 1, "event": "build_completed", "timestamp": "2026-04-27T21:27:33.141571Z"}
┃
▣ Build · glm-5.1
Man this has been a long time in the making 🤣 pynixd running in Kubernetes and creating backends dynamically, end-to-end functionality!!!!! (The Kubernetes logic is really shit)
| 21:32:10 |
| 31 Jan 2024 |
| Rickard Nilsson joined the room. | 22:05:29 |
lillecarl | Damn, I forgot to ask to take a picture w/ everyone. Thanks for a great evening everyone! 😁 | 23:37:57 |
| 1 Feb 2024 |
Claes | Yes great to meet everyone! Attendance was above all expectations! | 07:02:00 |
markuskowa | Thanks to everyone! Yesterday was absolutely great. | 09:27:57 |
| igorramazanov joined the room. | 10:09:05 |
| igorramazanov set a profile picture. | 10:10:05 |
igorramazanov | Yeah :) (hi all 👋) | 10:12:40 |
| 2 Feb 2024 |
lillecarl | { lib
, ...
}:
# $FLAKE should be configured to where your repository is located
# Preferably using direnv, you can also set it into your sessionVariables
let
dot_path = "${builtins.getEnv("FLAKE")}/sub/dir/";
dot_paths = lib.filesystem.listFilesRecursive dot_path;
dot_strings = builtins.map (x: builtins.toString x) dot_paths;
dot_prefixDeleted = builtins.map (x: builtins.replaceStrings [ dot_path ] [ "" ] x) dot_strings;
dotfile_outOfStoreLinked = lib.attrsets.genAttrs dot_prefixDeleted (name: {
source = config.lib.file.mkOutOfStoreSymlink (dot_path + name);
});
in
{
home.file = dotfile_outOfStoreLinked;
}
I know I mentioned linking things from $FLAKE repo into $HOME which I'm quite happy with. This is the implementation I'm using. It'll replicate a tree subdir into $HOME for you, nothing using home.file.x.recursive so you won't get collisions with folders and stuff 😄
| 20:18:15 |
| 3 Feb 2024 |
Claes | I didn't know you could do that! | 22:36:07 |
| 6 Feb 2024 |
igorramazanov | Would be nice to have an additional option: "does not really matter" and also allow multichoice | 19:51:29 |
| 7 Feb 2024 |
markuskowa | In reply to @igorramazanov:matrix.org Would be nice to have an additional option: "does not really matter" and also allow multichoice Will do it in the next iteration. I'm still exploring the options that matrix provides. | 17:48:01 |
| 12 Feb 2024 |
markuskowa | It looks like there is a slight preference for Wednesday. Should we go with Wednesday the 21st? | 09:58:19 |
| 13 Feb 2024 |
markuskowa | Ok, let's have the next meetup, on Wednesday 21.02. Same place, same time (19:00 at the Nordic light hotel). | 21:02:15 |
| 21 Feb 2024 |
markuskowa | See you tonight. | 09:47:33 |
Claes | Hi, I told some of you about my home audio setup. I created a gist with most of it here
https://gist.github.com/claes/61d0579df00a205385b5bd3ed235132c | 21:52:31 |
| 22 Feb 2024 |
kraem | In reply to @lillecarl:matrix.org
{ lib
, ...
}:
# $FLAKE should be configured to where your repository is located
# Preferably using direnv, you can also set it into your sessionVariables
let
dot_path = "${builtins.getEnv("FLAKE")}/sub/dir/";
dot_paths = lib.filesystem.listFilesRecursive dot_path;
dot_strings = builtins.map (x: builtins.toString x) dot_paths;
dot_prefixDeleted = builtins.map (x: builtins.replaceStrings [ dot_path ] [ "" ] x) dot_strings;
dotfile_outOfStoreLinked = lib.attrsets.genAttrs dot_prefixDeleted (name: {
source = config.lib.file.mkOutOfStoreSymlink (dot_path + name);
});
in
{
home.file = dotfile_outOfStoreLinked;
}
I know I mentioned linking things from $FLAKE repo into $HOME which I'm quite happy with. This is the implementation I'm using. It'll replicate a tree subdir into $HOME for you, nothing using home.file.x.recursive so you won't get collisions with folders and stuff 😄
ah this was what you were talking about yesterday! i'm gonna try to roll with some variant of this for a while. long time ago any of my systems were impure hehe | 19:05:22 |
kraem | LilleCarl (Salsa9): also did you mean https://github.com/PRQL/prql ? in that case we thought about the same project | 19:06:26 |
lillecarl | In reply to @kraem:ne.bul.ae ah this was what you were talking about yesterday! i'm gonna try to roll with some variant of this for a while. long time ago any of my systems were impure hehe Yeah there's the fileset lib I didn't even know about until i saw "umport" that recursively imports modules 😄 So it can be "nicer" | 19:07:49 |
lillecarl | In reply to @kraem:ne.bul.ae LilleCarl (Salsa9): also did you mean https://github.com/PRQL/prql ? in that case we thought about the same project Yeap, along with for running prql within pg https://github.com/kaspermarstal/plprql | 19:09:19 |
lillecarl | It's early work, but it's nice to see either way. SQL has had it's best days | 19:10:06 |
kraem | yeah i saw that, nice! last time i checked prql i think they were still talking about implementing the postgres layer | 19:10:12 |
kraem | this (https://github.com/neondatabase/neon) is pretty interesting as well, although not the same thing at all. it basically swaps out the storage layer of postgres, making it distributed and the backing storage is s3 :) | 19:24:28 |
| 26 Feb 2024 |
markuskowa | Wednesday seems to be a weekday that works, based on the last two meetups. How does Wednesday, March 27th work?
igorramazanov Do you know how to open a multiple-choice poll? In Element, it appears, I can only open single-choice polls.
| 18:09:44 |
igorramazanov | In reply to @markuskowa:matrix.org
Wednesday seems to be a weekday that works, based on the last two meetups. How does Wednesday, March 27th work?
igorramazanov Do you know how to open a multiple-choice poll? In Element, it appears, I can only open single-choice polls.
Let me check. Otherwise, we could, run an online poll, perhaps | 18:11:01 |
igorramazanov | The Matrix spec MSC3381 allow multichoice: https://github.com/matrix-org/matrix-spec-proposals/pull/3381/commits/211f5918189f4cf8ab7eb90250a3ba57c8212ae4 | 18:17:44 |
igorramazanov | * The Matrix spec MSC3381 allows multichoice: https://github.com/matrix-org/matrix-spec-proposals/pull/3381/commits/211f5918189f4cf8ab7eb90250a3ba57c8212ae4 | 18:17:53 |
igorramazanov | But it's not implement for Element :) | 18:27:29 |
igorramazanov | It's possible to create a multichoice poll in Discourse, done: https://discourse.nixos.org/t/nix-meetup-in-stockholm-sweden/32867/50 | 18:42:55 |
igorramazanov | markuskowa: | 18:43:04 |