| 12 May 2023 |
|  Federico Damián Schonborn changed their profile picture. | 00:58:18 |
| Federico Damián Schonborn changed their profile picture. | 19:34:06 |
| Federico Damián Schonborn changed their display name from Federico Schonborn to Federico Damián Schonborn. | 19:34:53 |
| 13 May 2023 |
 abathur | small update on the hydra jobs for the experimental installer
| 16:30:08 |
| 23 May 2023 |
|  K900 joined the room. | 09:32:58 |
| K900 | Hey, I fiiled https://github.com/DeterminateSystems/nix-installer/issues/474 already but was originally going to post here | 09:33:24 |
|  @ulrikstrid:matrix.org joined the room. | 09:52:59 |
 hexa | wait, steam does special handling of /nix already? | 10:23:31 |
| hexa | * wait, steam does special handling of /nix? | 10:24:34 |
 Zhaofeng Li | Not Steam but SteamOS. They just added a bind-mount of /nix since the root is read-only | 10:25:31 |
| hexa | basically what I meant 😄 | 10:25:53 |
| Zhaofeng Li | Pretty surprised that they added this though. Looks like they are somewhat interested in Nix to add such a specific workaround | 10:27:35 |
| K900 | There's a lot of people on youtube recommending people use Nix to install stuff on Steam Deck | 11:43:22 |
 graham | oh cool :D | 11:57:34 |
| 24 May 2023 |
 Ana | https://meet.google.com/dwy-yvjd-zpc | 14:00:00 |
| Ana | Meeting today | 14:00:05 |
 Matthew Kenigsberg | Théophane: could we get some decisions from the Nix team about which features need to be changed to start distributing the experimental installer? The differences are telemetry as well as what's listed here: https://github.com/NixOS/experimental-nix-installer#installation-differences | 14:41:51 |
| 25 May 2023 |
|  raitobezarius changed their display name from raitobezarius to disko in NixOS 23.11 when. | 13:32:56 |
| raitobezarius changed their display name from disko in NixOS 23.11 when to raitobezarius. | 13:37:48 |
| graham | I posted some performance statistics from github actions: https://discourse.nixos.org/t/introducing-the-determinate-nix-installer/25848/9 | 17:31:16 |
| raitobezarius |
That introduces the use of auto-uid-allocation, and the performance impact on Linux is stunning, making 100% of installs complete in fewer than 5s:
more of this please
| 17:32:05 |
| raitobezarius | beautiful | 17:32:10 |
| graham | we only serve 0.9.0 5% of the time right now, but we're about to increase that to 20% today | 17:32:41 |
| K900 | So was the slow part just shelling out to adduser however many times? | 17:32:53 |
| graham | (I'm quite cautious about auto-uid-allocation having surprising side effects) | 17:32:56 |
| graham | well, we exec it, ... but yeah | 17:33:05 |
| graham | * well, I think we exec it, ... but yeah | 17:33:08 |
| K900 | Cool | 17:33:33 |
| raitobezarius | In reply to @graham:grahamc.com
(I'm quite cautious about auto-uid-allocation having surprising side effects) Do you have HPC / big NFS deployment users? | 17:35:22 |
| raitobezarius | I'd target those for those features because they failed a lot with the old script | 17:35:29 |
