| 31 May 2026 |
hexa | in the meantime I'm wondering where we could get another aarch64-linux builder | 15:23:24 |
winston | i'll take a look at gnome | 15:30:09 |
leona | leaseweb has some (I don't know if they satisfy your requirements tho) | 15:31:14 |
leona | FRA (DE) / AMS (NL) if EU-based | 15:31:43 |
hexa | from 274 EUR + VAT | 15:36:35 |
hexa | ~331,50 EUR | 15:36:49 |
hexa | Redacted or Malformed Event | 15:36:53 |
hexa | with 128GB RAM 🫠| 15:37:04 |
hexa | 391 EUR with 256GB RAM | 15:37:42 |
hexa | Redacted or Malformed Event | 15:37:45 |
hexa | Redacted or Malformed Event | 15:37:48 |
winston | I ran into one build error so far https://github.com/NixOS/nixpkgs/blob/3109eaae18e09d0b8aef23dc2579e7d94b8d4b4e/pkgs/os-specific/linux/fuse/common.nix#L35-L39 shouldn't this be using fetchpatch2 because of the long commit hash format? | 17:17:45 |
winston | * I ran into one build error so far https://github.com/NixOS/nixpkgs/blob/90a6e9557b833557b9573503e2a980b3cfc57d41/pkgs/os-specific/linux/fuse/common.nix#L38-L42
shouldn't this be using fetchpatch2 because of the long commit hash format? | 17:18:20 |
ElvishJerricco | winston: open the link, you'll see short hashes | 17:19:35 |
winston | ah my bad I was looking at the from part in the first line yeah | 17:20:01 |
winston | then i guess its just a mismatch now for some reason | 17:20:10 |
Alyssa Ross | that's not good… | 17:21:29 |
winston | I guess they got the hash from nix store prefetch-url rather than fetchpatch | 17:21:52 |
Alyssa Ross | Changed two weeks ago… | 17:22:49 |
leona | 1w4j6f1awjrycycpvmlv0x5v9gprllh4dnbjxl4dyl2jgbkaw6pa was converted to sha256-UsLE14P50Ell0xo3PESk5tIHH2HLalsaWiEubO9KdPE=, but
λ nix hash to-sri 1w4j6f1awjrycycpvmlv0x5v9gprllh4dnbjxl4dyl2jgbkaw6pa --type sha256
sha256-6hqu5npSUN8I7XLZRiCl+b60Sweb1n2ZZz5LroIzkvA=
| 17:23:19 |
winston | yes thats what i get | 17:23:35 |
winston | and UsLE is what prefetch-url gave me, thus the assumption | 17:23:50 |
winston | I'll open a PR | 17:23:56 |
winston | https://github.com/NixOS/nixpkgs/pull/526429 | 17:26:07 |
| 1 Jun 2026 |
winston | finally hit the dependency cycle in sysprof | 01:10:11 |
winston | looks like sysprof-50.0-lib/libexec/sysprofd gets wrapped on staging, while it doesn't on master | 01:11:08 |
winston | and the wrapper references $out from $lib | 01:11:24 |
winston | I don't quite get why since I don't see a diff between staging and master? 🤔 | 01:13:27 |
winston | maybe its because of this? https://github.com/nixOS/nixpkgs/commit/a0ee89704420278abb4c28c7b974500b0d19d39c | 01:24:43 |
winston | yea I can confirm, reverting this makes sysprof build again | 01:34:33 |