| 22 Apr 2023 |
 hexa | cole-h: when do you want to be reminded? 😄 | 18:13:04 |
| 1 May 2023 |
 Emily | In reply to @cole-h:matrix.org
Remind me to take a look this weekend / next week.
I feel like it is something for the NixOS foundation, or at the very least someone on the ofborg team; I would feel very weird if someone not on the ofborg team / related to NixOS infra applied... (no offense) aarch64-darwin build queue is on the rise again (currently at 60).
cole-h friendly Scaleway OSP ping :^) | 16:04:31 |
 entheogenesis | 
Download Screenshot 2023-05-01 at 19.42.01.png | 23:42:10 |
| entheogenesis | Is OfBorg having trouble? | 23:42:14 |
| entheogenesis | 
Download Screenshot 2023-05-01 at 19.52.46.png | 23:52:54 |
| 2 May 2023 |
|  @emilytrau:matrix.org joined the room. | 10:19:59 |
| hexa | yes, too few aarch64-darwin builders | 12:05:09 |
| 4 May 2023 |
 raitobezarius | cole-h: do you know if I can run ofborg on a commit or delta locally? | 17:42:20 |
 cole-h | Depends on what you want to run | 17:42:33 |
| raitobezarius | I want to compute impacted maintainers | 17:42:43 |
| raitobezarius | also would be interested into bringing the a{1,2,3,...} syntax for globbing in commit messages | 17:43:31 |
| raitobezarius | hm I guess I would need to bring nixos modules support in the play to be able to parse the commit messages adequately | 17:46:15 |
| cole-h | Look at this: https://github.com/NixOS/ofborg/blob/released/ofborg/src/maintainers.nix | 18:00:27 |
| raitobezarius | Yeah I read it | 18:00:38 |
| raitobezarius | But I don't understand well the format of the input JSON | 18:00:43 |
| cole-h | https://github.com/NixOS/ofborg/blob/0f34038feb9b0ae9959c865608700c91d57b2590/ofborg/src/maintainers.rs#L51
It's a json list of file paths and separate attrpaths I believe | 18:02:00 |
| cole-h | ("separate" as in pkgs.hello.world -> ["pkgs" "hello" "world"]) | 18:02:14 |
| raitobezarius | thanks | 20:56:45 |
| raitobezarius | I guess it won't be enough for modules | 20:56:49 |
| raitobezarius | Because they're not really attrspaths | 20:56:59 |
| raitobezarius | Right now, what I'm doing is that I am adding touched_modules to track nixos/activation: ... or stuff like that | 20:57:15 |
| raitobezarius | then transform touched modules into a proper list of affected modules ideally | 20:57:25 |
| raitobezarius | by checking also the changed paths | 20:57:31 |
| raitobezarius | the most annoying thing though is it's unclear I can do ./test-pr $pr-number and run the whole nixpkgs eval logic on it without setting everything | 20:58:05 |
| raitobezarius | I'm wondering how hackable a thin wrapper to create this test-pr would require given the presence of a full fledged AMQP broker in this thing | 20:58:22 |
| raitobezarius | the code is full of easter eggs | 21:00:22 |
| raitobezarius | // FIXME: remove with rust/cargo update
#[allow(clippy::cognitive_complexity)]
| 21:01:58 |
| raitobezarius | I wonder if the easiest thing is to extract the logic that makes sense in a general way into a separate crate and test it then rewire it in ofborg | 21:03:09 |
| hexa | raitobezarius: you did see the call that was posted into the related issue? | 21:38:18 |
| hexa |
nix-instantiate --eval '<nixpkgs/nixos>' --arg configuration '{}' -A config.meta.maintainers --strict --json
| 21:38:35 |
