| 21 Jun 2026 |
 emily | esp since most everything that isn't already done can be done mid-cycle | 13:42:15 |
| 22 Jun 2026 |
 Grimmauld (any/all) | Will anyone scream if i just merge the gstreamer things? Those being:
https://github.com/NixOS/nixpkgs/pull/514056
https://github.com/NixOS/nixpkgs/pull/459808
I am not quite familiar enough with gstreamer to properly review this, but as it stands it isn't maintained at all and we are accumulating security issues. I would prefer just merging an update that at least on the surface looks reasonable and figure out the breakages later if there end up being problems. | 12:07:22 |
 K900 | https://github.com/NixOS/nixpkgs/pull/534242/changes can we get this into this cycle | 12:46:02 |
 Vladimír Čunát | I'd like to have the following staging-next done like this:
https://github.com/NixOS/nixpkgs/pull/534248 | 12:59:14 |
| Vladimír Čunát | (i.e. a quick one, utilizing staging-next-26.05 binaries) | 12:59:35 |
| Vladimír Čunát | I think we have about a day before we need to decide/start. Hydra is now a bit recovering after queue-runner changes. | 13:00:34 |
| K900 | I want to see a full staging run with the new queue runner | 13:00:48 |
| K900 | For science | 13:00:50 |
| Vladimír Čunát | I was more concerned about security fixes waiting since May 8th. | 13:03:55 |
| Vladimír Čunát | (those which were able to merge directly due to too many rebuilds) | 13:04:31 |
| Vladimír Čunát | * (those which were unable to merge directly due to too many rebuilds) | 13:04:37 |
 Alyssa Ross | how quick do you think it would be? | 13:37:49 |
| Vladimír Čunát | The new queue-runner situation is a bit of wild card here. | 13:39:47 |
 hexa | the time to integrate changes from staging into release has come down quite a bit | 13:40:19 |
| hexa | from days to hours | 13:40:23 |
| Vladimír Čunát | But assuming the old speed, I expect ~4 days instead of two weeks. | 13:40:23 |
| Vladimír Čunát | Yes, the merge lag. | 13:40:43 |
| Vladimír Čunát | Though the lag to catch up with NixOS tests seems to remain, at least to some extent. | 13:40:55 |
| Vladimír Čunát | (or whatever is slowing it down at this very moment) | 13:41:06 |
| hexa | if we only schedule 4 steps in 30s that is bad for big closures with many small files | 13:41:21 |
| Vladimír Čunát | * (or whatever is slowing it down at this very moment, e.g. on nixos-26.05) | 13:41:23 |
| hexa | Redacted or Malformed Event | 13:41:29 |
| Vladimír Čunát | * (or whatever is slowing it down at this very moment, e.g. on https://hydra.nixos.org/jobset/nixos/release-26.05 ) | 13:41:52 |
| Vladimír Čunát | Still, "Steps completed per minute" metric shows much higher than with the old queue-runner. | 13:42:53 |
| Vladimír Čunát | So even this phase is probably significantly faster than it used to be. | 13:43:28 |
| Alyssa Ross | have we still not disabled the non-blocker NixOS tests? | 13:44:10 |
| Grimmauld (any/all) | also don't forget we recently killed half the nixos tests | 13:44:16 |
| Vladimír Čunát | We haven't. I've had too many other higher-priority tasks so far. | 13:44:44 |
| Grimmauld (any/all) | I meant killing the all-drivers tests (https://github.com/NixOS/nixpkgs/pull/532396) | 13:45:42 |
| Vladimír Čunát | Yes, I know, but I don't expect it helped dramatically with building all those etc and disk-image derivations. | 13:46:32 |
