!9IQChSjwSHXPPWTa:lix.systems

Lix

961 Members
Lix user channel. Feel free to discuss on-topic issues here and give each other help. For matrix.to links to the rest of the Lix channels, see: https://wiki.lix.systems/books/lix-organisation/page/matrix-rooms259 Servers

Load older messages


SenderMessageTime
10 Jul 2025
@charles:computer.surgeryCharles probably just man nixos-install and then use the --flake option and any other options as you see fit 05:16:24
@charles:computer.surgeryCharlesgiven that you seem to be using flakes05:16:49
@arianvp:matrix.orgArianWould be nice if you could connect to a remote builder or daemon store over vsock directly. Is this something that is supported?08:43:48
@raitobezarius:matrix.orgraitobezarius
In reply to @arianvp:matrix.org
Would be nice if you could connect to a remote builder or daemon store over vsock directly. Is this something that is supported?
Explain the scenario
09:47:07
@molly:161.rocksMollyfollowing up on this – i didn't explain myself particularly well in my first message, but i'm referencing this PR on the nix-community flake-compat: https://github.com/nix-community/flake-compat/pull/4. the problem statement: i'm importing the flake in https://github.com/cachix/nixpkgs-python into my configuration using flake-compat because i need a python 3.7 interpreter for dealing with some appliances at work. the nixpkgs-python flake has its own pin of nixpkgs, and importing the flake will cause a copy of this pinned nixpkgs to be downloaded and evaluated. what i'd like to be able to do is externally override this input in the nixpkgs-python flake to use the copy of nixpkgs which i use to build the rest of the configuration on my work machine (to cut down on the eval time and number of nixpkgs copies sitting in my store).10:47:05
@molly:161.rocksMollycurrently i have flake-compat pinned in my configuration to the commit from that PR, but i thought i might as well ask about it in here in case there's interest in similar functionality in the lix flake-compat10:48:19
@raitobezarius:matrix.orgraitobezariushmmmmmmmmmm, that sounds like a cool idea imho15:29:18
@raitobezarius:matrix.orgraitobezarius Molly would you open an issue in lix's flakes-compat with that feature and provide your design, etc.? 15:29:55
@raitobezarius:matrix.orgraitobezariusthis way, jade & co. can take care of it and say they would accept a PR for it15:30:02
@molly:161.rocksMollyi can't really offer any suggestions for design other than point at the other PR, but sure, i can open an issue for the feature15:31:01
@molly:161.rocksMolly done 16:00:28
@raitobezarius:matrix.orgraitobezariusthank you!16:41:17
11 Jul 2025
@emma:rory.gayEmma [it/its]what in hell is going on here?10:45:24
@emma:rory.gayEmma [it/its]clipboard.png
Download clipboard.png
10:45:26
@emma:rory.gayEmma [it/its] i was not expecting to see lix using 9GB of RAM to do a nix run nixpkgs#gittyup 10:45:46
@emma:rory.gayEmma [it/its] nix (Lix, like Nix) 2.91.1 10:47:04
@k900:0upti.meK900Try latest?10:47:37
@emma:rory.gayEmma [it/its]ah, seems it was actually running gittyup but didnt rename the executable according to btop10:47:46
@emma:rory.gayEmma [it/its] [2247546.742048] Out of memory: Killed process 3334734 (.gittyup-wrappe) total-vm:39094832kB, anon-rss:7048740kB, file-rss:596kB, shmem-rss:0kB, UID:1000 pgtables:72920kB oom_score_adj:0 10:48:08
@emma:rory.gayEmma [it/its]not a lix issue, just weird btop behavior compounded with gittyup for some reason very much disliking this repo im working in10:50:05
@raitobezarius:matrix.orgraitobezarius
In reply to @emma:rory.gay
i was not expecting to see lix using 9GB of RAM to do a nix run nixpkgs#gittyup
sounds like just average nixpkgs behavior
13:54:12
@emilazy:matrix.orgemily9 GiB for eval of a package and not a NixOS system configuration? sounds excessive.14:04:55
@emilazy:matrix.orgemily(and seemingly not what was going on thankfully)14:05:07
@emma:rory.gayEmma [it/its]yeah it wasnt a nixpkgs issue apparently14:33:36
@emma:rory.gayEmma [it/its]jut weird exec behavior14:33:46
@aloisw:julia0815.dealoiswThe process name should change on exec, but some system monitors don't notice because they cache it to speed things up.15:49:09
@raitobezarius:matrix.orgraitobezarius
In reply to @emilazy:matrix.org
9 GiB for eval of a package and not a NixOS system configuration? sounds excessive.
new CLI feels like it's doing a lot of unnecessary nixpkgs evals
18:09:45
@emilazy:matrix.orgemily for nix run? that would surprise me 18:11:07
@raitobezarius:matrix.orgraitobezariusit's just my own bias towards new CLI18:13:25
@edersohe:matrix.orgEder Sosa joined the room.20:13:34

There are no newer messages yet.


Back to Room ListRoom Version: 10