| 27 Mar 2023 |
 nbp | If you do not plan on solving constraints, then you are going to have assertions or conflicting configuration which are left to the user to resolve. | 17:19:23 |
| 29 Mar 2023 |
|  prescientmoon joined the room. | 13:25:50 |
 snowytrees | Reading nbp 's comment (https://github.com/NixOS/rfcs/pull/3#issuecomment-370151582) in the RFC after eelco introduced his configurations idea (https://gist.github.com/edolstra/29ce9d8ea399b703a7023073b0dbc00d) cleared up a lot for me how you can build on SOS. Dropping the link because it was pretty deep in the comments | 15:22:36 |
| 2 Apr 2023 |
 tea | Has something like doing nixpkgs-review on github actions been tried out yet? | 14:41:59 |
| tea | would speed up time to merge and complexity by a lot | 14:42:10 |
 Alyssa Ross | there's nowhere near enough capacity, is there? | 14:42:28 |
| Alyssa Ross | i feel like somebody was talking about this in another room the other day | 14:42:45 |
| tea | In reply to @qyliss:fairydust.space
there's nowhere near enough capacity, is there? capacity =? | 14:43:08 |
| Alyssa Ross | wait it was you who was having this conversation | 14:43:25 |
| tea | where? | 14:44:08 |
| Alyssa Ross | in #tvl | 14:44:19 |
| Alyssa Ross | 2023-03-26 | 14:44:24 |
| tea | this is for nixpkgs-review and not nixpkgs as a whole | 14:44:28 |
| Alyssa Ross | assuming you're tea|m | 14:44:28 |
| tea | yes | 14:44:32 |
| tea | there is a limit of 20 concurrent jobs which would get really jammed | 14:44:48 |
| tea | so maybe we'd need to filter somehow | 14:44:55 |
| Alyssa Ross | yeah | 14:45:06 |
| Alyssa Ross | nixpkgs-review is often huge | 14:45:11 |
| Alyssa Ross | so it would be useful for the subset of packages that only have one or two things that depend on them | 14:46:08 |
| Alyssa Ross | but I feel like there aren't very many of those? | 14:46:24 |
| Alyssa Ross | usually stuff doesn't have any dependents, or it has a bunch | 14:46:25 |
 K900 | It's a very bimodal distribution yeah | 14:46:43 |
| K900 | It's either 1 or the entire world | 14:46:49 |
| K900 | And very little in between | 14:46:52 |
| Alyssa Ross | ofborg takes care of the 1 case, and the staging process takes care of the entire world case | 14:47:08 |
 @piegames:matrix.org | In reply to @k900:0upti.me
It's either 1 or the entire world 1, 2, many :) | 14:47:51 |
| Alyssa Ross | (personally I don't find nixpkgs-review hugely useful anyway, because it doesn't tell you whether stuff was already broken before, so it needs a bunch of post-hoc investigation anyway) | 14:48:28 |
| Alyssa Ross | unless you get lucky and everything builds | 14:48:47 |
| Alyssa Ross | how much RAM do you get on GitHub actions? | 14:49:14 |
