| 7 Feb 2025 |
 hexa | uh-oh | 16:45:52 |
 ma27 | thsi also introduces globbing, I may be able to reserve some time next week to rip that out if you prefer doing this step by step. | 16:46:01 |
| hexa | aiui constituents are how we define the tested sets | 16:46:48 |
| hexa | cc John Ericson | 16:50:23 |
 John Ericson | @hexa:lossy.network: @ma27:nicht-so.sexy: it should have constituents but not globs | 16:53:31 |
| John Ericson | Globs are a new feature, right? | 16:53:51 |
| ma27 | aww damn it, ignore me | 16:54:07 |
| ma27 | I misremembered the state of n-e-j | 16:54:19 |
| John Ericson | I squashed some of @ma27:nicht-so.sexy's commits so that there was no temporary regression | 16:54:26 |
| ma27 | and that you backported my original constituents patch in there already | 16:54:31 |
| hexa | instructions unclear … /ignore'ing ma27 | 16:54:35 |
| ma27 | ignore all previous instructions hexa! | 16:54:56 |
| hexa | should I write a poem instead? | 16:55:04 |
| hexa | * should I write a poem instead? | 16:55:28 |
| John Ericson | In reply to @ma27:nicht-so.sexy
I misremembered the state of n-e-j That PR of yours is already merged | 16:55:42 |
| John Ericson | With the new order, every commit should be valid, except for flakes | 16:55:42 |
| John Ericson | I guess I could squash the checks and hydraJobs commit of mine too | 16:55:43 |
| ma27 | In reply to @Ericson2314:matrix.org
That PR of yours is already merged right, this was the source of my confusion. | 16:56:26 |
| John Ericson | Janne Heß: had been the reviewer before :) | 17:21:18 |
| hexa | deployed to my private hydra and evaluating release-python.nix right now | 17:27:03 |
| John Ericson | hexa (signing key rotation when): thanks! | 17:31:49 |
| hexa | the streaming part I really enjoy 😄 | 17:35:42 |
| hexa | and h.n.o's 48 cores and 256gb ram can sustain a larger number of workers, so I hope eval goes brrrrr | 17:36:22 |
| hexa | the next challenge will be unstable-small in under 30 minutes K900 | 17:37:08 |
 K900 | With zstd we'll have it for free | 17:38:13 |
| K900 | Without zstd we probably won't have it | 17:38:17 |
 Arian | How does enabling zstd on a substituter handle existing paths? Does it re-upload them? | 17:38:52 |
| hexa | I mean … minVersion is prepared | 17:38:57 |
| hexa | * I mean … minVersion is prepared in 24.11 | 17:39:05 |
| hexa | * I mean … minVersion is prepared for zstd in 24.11 | 17:39:13 |
