| 7 Jun 2026 |
| Likius Inik joined the room. | 05:20:06 |
| lnk3 joined the room. | 10:11:53 |
| lnk3 changed their display name from Luca Ruperto to lnk3. | 10:12:20 |
lnk3 | Is passthru.updateScript = nix-update-script {}; redundant?
I was adding it to obs-cmd some time ago to bump it to 1.0.0, but now I noticed the package updated without it:
$ git log -1 -- pkgs/by-name/ob/obs-cmd/package.nix
commit 780c676b733629966efeb4bbaf5a5d88d5b3d23d
Author: R. RyanTM <ryantm-bot@ryantm.com>
Date: Mon Apr 13 19:26:51 2026 +0000
obs-cmd: 0.31.3 -> 1.0.0
The bot did the update, but the package doesn't explicit the script:
$ git diff --staged
diff --git a/pkgs/by-name/ob/obs-cmd/package.nix b/pkgs/by-name/ob/obs-cmd/package.nix
index 9e11f3d97c1d..c9b916e2ac93 100644
--- a/pkgs/by-name/ob/obs-cmd/package.nix
+++ b/pkgs/by-name/ob/obs-cmd/package.nix
@@ -2,6 +2,7 @@
+ nix-update-script,
}:
+ passthru.updateScript = nix-update-script {};
+
By reading the docs it seems that there is a "chain" it follows:
It uses package repository information from Repology.org (though that requires other projects to detect the update), the GitHub releases API, and, if available, the package passthru.updateScript
I remember when I added the change I ran nix-update -A obs-cmd and it updated successfully, even if the bot logs didn't yet pick up the new version.
So maybe it is useful to add it (?)
| 10:15:44 |
Austin Horstman | IIRC it was necessary to get around a local maintainer update script logic so that you can use an in repo script to update a plugin by name or a set of plugins for a maintainer | 18:42:28 |
Austin Horstman | ie num = "nix-shell maintainers/scripts/update.nix --argstr maintainer $1"; | 18:43:37 |
Austin Horstman | maybe related to get-script ? pkg: pkg.updateScript or null, | 18:44:54 |
| 9 Jun 2026 |
| coolcuber set a profile picture. | 19:10:53 |
| 10 Jun 2026 |
lnk3 | Let's see if I understand what you are saying: passthru.updateScript is needed for maintainers to have a place to write scripts for their packages update. Each maintainer can run a command passing its name and the get-script function will run for each of the maintainers modules and run the script if present.
So regarding my question I can therefore assume that the bot doesn't really care about the presence of the updateScript. It has scripts of its own to update packages generically. If an updateScript it's not strictly required I won't add it since the bot should be able to update it on its own. I can maybe open a PR if I want to speed up a package update since the bot may have a long queue before reaching the package.
Also if I specify passthru.updateScript for packages that "aren't mine" then the maintainer will have them in their nix-shell maintainers/scripts/update.nix --argstr maintainer NOTME output when they didn't before and they might not be pleased.
Am I wrong?
| 13:13:00 |
lnk3 | Yep just tried it. Adding updateScript may not be enjoyed by the actual maintainer if they use the script
$ grep updateScript pkgs/by-name/ob/obs-cmd/package.nix; nix-shell maintainers/scripts/update.nix --argstr maintainer ramblurr
Going to be running update for following packages:
- davis-5.4.2
- gost-3.2.6
- hacompanion-1.0.30
- microsocks-1.0.5
- ocis_5-bin-5.0.9
- roon-server-2.67.1661
Press Enter key to continue...^C
obs-cmd not present above. Now I add the updateScript part and it will be present:
$ grep update pkgs/by-name/ob/obs-cmd/package.nix; nix-shell maintainers/scripts/update.nix --argstr maintainer ramblurr
passthru.updateScript = nix-update-script {};
nix-update-script,
passthru.updateScript = nix-update-script {};
Going to be running update for following packages:
- davis-5.4.2
- gost-3.2.6
- hacompanion-1.0.30
- microsocks-1.0.5
- obs-cmd-1.0.0
- ocis_5-bin-5.0.9
- roon-server-2.67.1661
Press Enter key to continue...^C
| 13:19:30 |
| 19 Sep 2023 |
| evils joined the room. | 16:15:20 |
ryantm | raitobezarius: No, not really. It's been an idea for a long time to have grouped package updates like that, but it hasn't been implemented. | 16:52:21 |
evils | hi, second link in the channel topic is dead | 19:21:58 |
evils | is there a way to see what packages the bot tried to update but failed on? | 19:22:23 |
Artturin | In reply to @evils:nixos.dev is there a way to see what packages the bot tried to update but failed on? https://r.ryantm.com/log/ | 19:23:49 |
Artturin | https://ryantm.github.io/nixpkgs-update/nixpkgs-maintainer-faq/#nixpkgs-maintainer-faq | 19:23:56 |
Artturin | * from https://ryantm.github.io/nixpkgs-update/nixpkgs-maintainer-faq/#nixpkgs-maintainer-faq | 19:23:58 |
evils |  Download image.png | 19:24:43 |
evils | oof that's a subtle link underline | 19:24:50 |
evils | (i was looking at the page :) | 19:25:02 |
Artturin | yeah | 19:26:15 |
evils | ok, why do all KiCad update attempts end in "Do not update GNOME during a release cycle"? | 19:30:21 |
ryantm |
| 19:44:27 |
ryantm | * [ryantm@replit1:~/p/nixpkgs]$ EDITOR=echo nix edit -f . kicad
/home/ryantm/p/nixpkgs/pkgs/applications/science/electronics/kicad/default.nix
[ryantm@replit1:~/p/nixpkgs]$ cat /home/ryantm/p/nixpkgs/pkgs/applications/science/electronics/kicad/default.nix | grep gnome
, gnome
"--prefix XDG_DATA_DIRS : ${gnome.adwaita-icon-theme}/share" | 19:44:32 |
Artturin | the check here is wrong https://github.com/ryantm/nixpkgs-update/blob/8fedf8f6f29280578e335bf4483c9f88d4bcdbaa/src/Skiplist.hs#L151 | 19:45:39 |
Artturin | just a naive grep | 19:45:49 |
Artturin | maybe a better way would be to check for teams.gnome.members | 19:47:29 |
Artturin | still has false positives but way less | 19:47:37 |
ryantm | https://github.com/ryantm/nixpkgs-update/commit/2fda8bcb73776c3e11c0155fa15e138a71e1f189 | 19:49:37 |
evils | In reply to @artturin:matrix.org the check here is wrong https://github.com/ryantm/nixpkgs-update/blob/8fedf8f6f29280578e335bf4483c9f88d4bcdbaa/src/Skiplist.hs#L151 huh, i see kicad a bit further down there, what does binariesStickAround imply? | 19:54:51 |