| 28 Oct 2025 |
 emily | it seems like a poor substitute for variant sets and nixpkgsFun | 18:21:57 |
| emily | it also… ignores its second argument?? | 18:23:13 |
| emily | I'm going to add a warning to it at least. | 18:23:18 |
 Vladimír Čunát | Yes, I think it won't. | 18:25:12 |
| emily | is making it use NIX_ABORT_ON_WARN or similar feasible, or would that break nix-eval-jobs? | 18:34:50 |
| emily | I suppose I should say abort-on-warn instead | 18:35:46 |
| emily | ok, https://github.com/NixOS/nixpkgs/pull/456527 | 18:56:48 |
| 29 Oct 2025 |
 K900 | Started a new eval on unstable-small to pick up the new kernels because it looks like there's security stuff in there | 14:39:29 |
 Grimmauld (any/all) | btw re kernel: it seems gentoo enables some ARCH_QCOM stuff for their default kernels, and we don't. Was struggling with a friend yesterday to make a nixos boot. It still doesn't boot, so we are still probably missing some options. With aarch64 laptops slowly gaining popularity, would there a point at which we too set those default enables? | 15:04:36 |
| Grimmauld (any/all) | * btw re kernel: it seems gentoo enables some ARCH_QCOM stuff for their default kernels, and we don't. Was struggling with a friend yesterday to make a nixos boot. It still doesn't boot, so we are still probably missing some options. With aarch64 laptops slowly gaining popularity, would there be a point at which we too set those default enables? | 15:04:58 |
| K900 | Honestly probably like now | 15:07:32 |
| K900 | If you know the exact options and they don't break other stuff | 15:07:42 |
| Grimmauld (any/all) | Alright then, will open a PR once we figure out what exactly is missing and we get that laptop booting | 15:09:16 |
 Alyssa Ross | Generally if other reputable distros enable config options that's a good sign that we should too. (There are obviously exceptions.) | 15:20:43 |
 dramforever | ARCH_QCOM is y in arm64 defconfig | 15:53:49 |
| dramforever | i don't think there's any other "ARCH_QCOM stuff" | 15:54:18 |
| dramforever | it's possible we're missing some bools | 15:59:22 |
| 30 Oct 2025 |
 Fabián Heredia | I would say the time is always right. The policy is around other distros already using values as default and enabling hardware seems really reasonable.
Some time ago I made some effort to better compartamentalize the option by ISA so aarch64 specific stuff could be enabled on that platform easily. | 05:33:08 |
| Fabián Heredia | ah wait nvm, it was only power-management options | 05:38:16 |
| Fabián Heredia | https://github.com/NixOS/nixpkgs/pull/165547/commits/cc8456effe723a6489b6f89fdafa7e877245742d
Anyways, same idea, // optionalAttrs (platform)
| 05:38:42 |
| Fabián Heredia | I should also do a cleanup now that 5.4 was removed | 05:41:44 |
 j-k | I was in the middle of asking about the jobs here:
https://hydra.nixos.org/job/nixpkgs/trunk/unstable#tabs-constituents
But it looks like someone is going through and cancelling some of the old pending work | 10:27:46 |
| j-k | aggregate job 311452680 I believe in you 🙏 | 10:28:29 |
| Vladimír Čunát | If only trunk wouldn't commonly generate 10k rebuilt jobs daily (roughly). | 10:29:35 |
| Vladimír Čunát | Most of cancellations that you see are from
https://hydra.nixos.org/admin/clear-queue-non-current | 10:30:29 |
| j-k | does merge trains now help with that a bit? | 10:30:53 |
| j-k | * do merge trains now help with that a bit? | 10:30:58 |
| j-k | So it's this one I should be watching for staging -> master?
https://hydra.nixos.org/jobset/nixpkgs/staging | 10:31:05 |
| Vladimír Čunát | Merge queue is unrelated, I believe. | 10:31:15 |
| Vladimír Čunát | * Merge queue is unrelated, I believe. | 10:31:59 |
