| 26 Nov 2025 |
 ToxicFrog | K900: I'm guessing the recent one where a bad config push caused them to crashloop for a while | 11:39:34 |
 truh | I've definitly deployed crashloops with nixos too. | 12:05:01 |
 Sofie 🏳️⚧️ (she/her) | @k900:0upti.me just that I remember to, may you leak something about the secret handling thing | 13:05:30 |
| Sofie 🏳️⚧️ (she/her) | * @k900:0upti.me just that I remember to, may you link to something about the secret handling thing | 13:05:39 |
| Sofie 🏳️⚧️ (she/her) | * @k900:0upti.me just that I remember to say so sorry ping, may you link to something about the secret handling thing | 13:05:49 |
 K900 | I don't think it's written down anywhere, but it has been discussed a hunch | 13:46:49 |
 aloisw | … or it doesn't try to write to the state directory, unlike the real store. | 15:28:02 |
|  deman9 joined the room. | 18:45:36 |
|  Ellie (The Fake One) changed their display name from ellie (the hot one) to Ellie (The Fake One). | 19:38:26 |
|  dantefromhell left the room. | 21:19:59 |
|  Milo joined the room. | 23:24:37 |
| 27 Nov 2025 |
 WeetHet | Why does nix repl follow the structure of nix-command while not being experimental? Was it always like that? Is this yet another cppnix historical artifact? Does this mean that nix-command is impossible to remove without a breaking change? | 12:44:01 |
| WeetHet | nix repl docs suggest that
│ Warning
│ This program is experimental and its interface is subject to change.
Name
nix repl - start an interactive environment for evaluating Nix expressions
Synopsis
nix repl [option...] installables...
Note: this command's interface is based heavily around installables, which
you may want to read about first (nix --help).
| 12:44:48 |
| WeetHet | Why is it not behing nix-command | 12:44:55 |
| WeetHet | Why is there no nix-repl that would work like a normal nix2 program? | 12:45:13 |
 toonn | It was always like this, there was no repl before nix repl. | 12:46:37 |
| WeetHet | Would making it dependent on nix-command being enabled and add a proper nix-repl, or would that be considered breaking? | 12:52:10 |
| toonn | What would be the benefit? Other than satisfying a desire for consistency? | 12:55:09 |
|  @sntx:matrix.org left the room. | 13:36:10 |
 goldstein | Atemu unflake is now public! https://discourse.nixos.org/t/unflake-flake-dependencies-for-non-flake-projects-and-a-way-to-stop-writing-follows/72611 / https://codeberg.org/goldstein/unflake | 18:46:41 |
| aloisw | In reply to @weethet:catgirl.cloud
Why does nix repl follow the structure of nix-command while not being experimental? Was it always like that? Is this yet another cppnix historical artifact? Does this mean that nix-command is impossible to remove without a breaking change? Yes, historical artifact. Actually there was a separate nix-repl at some point but it got merged into the tree. | 19:49:53 |
 vringar | Hey,
I'm currently trying to set up nixos on ZFS by following the nixos.wiki/wiki/ZFS guide from the NixOS installer image
However, the image doesn't have ZFS kernel module active, the /etc/nixos/configuration.nix is read-only, so I can't add it but zpool create only works with the kernel module present.
How can I best bootstrap myself into an environment with the ZFS kernel module present so I can configure the internal disk?
My first thoughts are remounting the root file system to be rw or to install NixOS to ext4 on the primary disk, then boot there, follow the tutorial from there and later nuke the ext4 partition and reintegrate it into the zpool | 19:45:18 |
 helle (just a stray cat girl) | I mean I kind of prefer the nix-command structure personally, but I haven't know nix without it | 20:00:17 |
 raitobezarius | The NixOS installer image has ZFS on the LTS kernel | 20:06:56 |
| raitobezarius | Simply reboot in the LTS kernel and you should have ZFS I think? | 20:07:03 |
| vringar | In reply to @raitobezarius:matrix.org
Simply reboot in the LTS kernel and you should have ZFS I think? Ahhh, my bad
Didn't consider that when choosing the installer image.
Thanks! | 20:12:03 |
| toonn | Is unflake flake-compat on steroids? | 23:12:51 |
| 28 Nov 2025 |
|  yqrashawn joined the room. | 00:35:31 |
 antifuchs | Oh neat, unflake looks like fun, but way different from flake-compat. The latter tries to work with the flake as written; unflake looks like it rewrites your source tree | 04:43:00 |
| goldstein | flake-compat works with flake.lock, so it uses built-in resolver
unflake doesn’t interact with flake.lock at all. it resolves dependencies internally, unifies them and produces its own lockfile (or integrates with npins for locking) | 11:04:01 |
