| 24 Feb 2026 |
 vcunat | Redacted or Malformed Event | 07:45:13 |
 Alyssa Ross | Wonder if that's automatable. Feels like it should be... | 07:45:47 |
 K900 | We can probably auto-merge staging-nixos into next? | 07:46:26 |
| K900 | Like instead of going master -> next -> staging, we can go master -> staging-nixos -> next -> staging | 07:46:42 |
|  makefu joined the room. | 13:02:48 |
| Alyssa Ross | Oh yeah perfect | 14:48:50 |
 tazjin | we're several versions behind on libxkbcommon (~ every GUI thing depends on it), I rebuilt the world and tested with this PR: https://github.com/NixOS/nixpkgs/pull/487689
no build failures across a bunch of heterogeneous nixos configs with lots of desktop software, and the systems I've tested run fine. I think it makes sense to move forward with this, as it's going to start blocking more and more updates (right now, for example, latest river depends on tihs)
| 14:57:30 |
| K900 | Sure, let's just get this in | 14:59:20 |
| makefu | hey all, currently calibre is broken on unstable and master ( https://hydra.nixos.org/build/322102655 ), last success was at https://hydra.nixos.org/build/321911294 . error message states it cannot find qt6 qmake anymore. i tried to pinpoint the issue (or have rather have mr computer trying to build the package until it is successful) and it came up with the following solution https://p.krebsco.de/1x23192 however i have absolutely no idea if this is a good idea and i had no plan to touch qt6 buildhooks in first place.
setting env.QMAKE="${qt6.qtbase}/bin/qmake" in calibre directly didnt help
before i create a PR, could anybody have a quick look at the solution above?
| 17:37:25 |
| K900 | I believe we have a fix for that in the next staging | 17:37:43 |
| makefu | is staging-next more "fresh" than master? | 17:38:26 |
| K900 | No | 17:39:36 |
| K900 | staging-next is currently behind master, as Hydra is building staging-25.11 | 17:39:45 |
| K900 | We should have the next staging-next cycle probably somewhere around the end of the week | 17:40:02 |
| makefu | the issue still exists in master when trying to build calibre | 17:52:33 |
| K900 | Yes, and the fix is in staging | 17:52:51 |
| K900 | Not staging-next | 17:52:52 |
| K900 | Which is ahead of master | 17:52:55 |
| makefu | ah!!! | 17:53:06 |
| K900 | https://github.com/NixOS/nixpkgs/blob/master/CONTRIBUTING.md#staging | 17:53:13 |
| makefu | okay thanks, that makes more sense | 17:53:13 |
| vcunat | It's unlikely to happen this week. | 17:55:59 |
| makefu | so when i am checking build failures in unstable because i did nix flake update nixpkgs and now my system does not build i should generally check both channels, staging and master if the package builds? or was calibre special that it needed fixing via the staging branch? | 17:59:49 |
| vcunat | staging tends to be hard to check. You'd be doing lots of building work. | 18:01:12 |
| vcunat | And it more often breaks things than fixes 😅 | 18:01:51 |
| K900 | Calibre wasn't special, but qmake in general was | 18:01:53 |
 WeetHet | You can always overlay the package to fix the build yourself (worst case use it from an earlier nixpkgs instance) and wait for the fix to propagate to unstable and realistically that's all you need to care about if you're not interested in contributing big changes that would touch staging | 18:04:10 |
 Fabián Heredia | In reply to @k900:0upti.me
Calibre wasn't special, but qmake in general was How many rebuilds is the qmake fix currently in staging? | 18:04:49 |
| K900 | About 5k | 18:09:03 |
| WeetHet | Did the effort to make stdenv non-cc by default die out? | 18:18:54 |
