| 13 Feb 2023 |
 cole-h | Maybe unrelated, but ofborg is having a bit of an issue right now. All evaluators are unable to do their job. Investigating. | 20:59:25 |
| cole-h | OK, they should be slowly coming back. Long story short: weirdness with the GitHub App RSA key (that allows ofborg to sign a JWT to do its thing) | 21:52:06 |
 Winter (she/her) | Huh, did it expire? | 21:52:21 |
| Winter (she/her) | Thank you nonetheless as always. | 21:52:28 |
| cole-h | No, I screwed something up :) | 21:52:30 |
| Winter (she/her) | * Huh, did it expire or something stupid? | 21:52:32 |
| cole-h | Fix: https://github.com/NixOS/ofborg/pull/634
I'll be deploying this tomorrow. | 22:48:14 |
| 15 Feb 2023 |
| Winter (she/her) | cole-h: Would you mind confirming that https://github.com/NixOS/nixpkgs/pull/216231 only had one build attempt per system?
I ask because the author is saying that they could have sworn it succeeded on OfBorg's x86_64-darwin when it timed out (plus, the current derivation fails on Darwin -- there's no chance it actually did succeed even if it didn't time out). Even though I don't think it's actually the case (as the logs site confirms), I just want to be 100% sure. Thank you! | 05:36:54 |
| cole-h | In reply to @winterqt:nixos.dev
cole-h: Would you mind confirming that https://github.com/NixOS/nixpkgs/pull/216231 only had one build attempt per system?
I ask because the author is saying that they could have sworn it succeeded on OfBorg's x86_64-darwin when it timed out (plus, the current derivation fails on Darwin -- there's no chance it actually did succeed even if it didn't time out). Even though I don't think it's actually the case (as the logs site confirms), I just want to be 100% sure. Thank you! Whatever you see on logs.nix.ci is the truth of what happened :P So x86_64-darwin only ever timed out: https://logs.nix.ci/?attempt_id=7e2121db-9d89-4c2b-85a7-8c0dc23fcccd&key=nixos%2Fnixpkgs.216231 | 15:10:52 |
| 17 Feb 2023 |
| Winter (she/her) | In reply to@cole-h:matrix.org
Whatever you see on logs.nix.ci is the truth of what happened :P So x86_64-darwin only ever timed out: https://logs.nix.ci/?attempt_id=7e2121db-9d89-4c2b-85a7-8c0dc23fcccd&key=nixos%2Fnixpkgs.216231 That's what I thought, thanks. | 02:26:03 |
| 21 Feb 2023 |
|  Yuu Yin joined the room. | 10:27:55 |
| Yuu Yin | is there an issue with ofborg-eval-darwin?
https://gist.githubusercontent.com/GrahamcOfBorg/0f47ada066839a60bf38126e0b4eff14/raw/51b5d34c48a34d8d80e9215f17f30ee1ce835ed4/ofborg-eval-darwin
from https://github.com/NixOS/nixpkgs/pull/217424
| 10:31:12 |
| cole-h | Just looks like it was a networking issue at some point. Couldn't reproduce it locally, so I just told ofborg to re-eval. | 15:10:31 |
| cole-h | (Yeah, re-eval succeeded.) | 15:39:13 |
 Linux Hackerman | Nix expressions or life choices? | 15:39:39 |
| Linux Hackerman | (scnr) | 15:39:41 |
| Yuu Yin left the room. | 23:49:42 |
| 22 Feb 2023 |
 Alyssa Ross | Does OfBorg include broken packages in rebuild counts? https://github.com/NixOS/nixpkgs/pull/217591#issuecomment-1440705382 | 19:49:53 |
| cole-h | It's possible: https://github.com/NixOS/ofborg/blob/cf2c6712bd7342406e799110e7cd465aa250cdca/ofborg/src/outpaths.nix#L22 | 19:54:46 |
| Alyssa Ross | Ahhh | 20:09:08 |
| cole-h | I'd have to do more research / testing to know for sure, but I think that would make sense | 20:09:38 |
| Alyssa Ross | I wonder if that should be dropped? AIUI broken doesn't block accessing any attributes, so it shouldn't miss any eval breakages to disallow broken. | 20:09:55 |
| cole-h | It comes from a commit entitled "Support checking meta" -- so maybe a broken package does weird things with meta? | 20:10:55 |
| cole-h | * It was introduced in a commit entitled "Support checking meta" -- so maybe a broken package does weird things with meta? | 20:11:05 |
| Winter (she/her) | In reply to @qyliss:fairydust.space
I wonder if that should be dropped? AIUI broken doesn't block accessing any attributes, so it shouldn't miss any eval breakages to disallow broken. Broken only stops you if you're evaluating every attribute AFAICT | 20:11:14 |
| Alyssa Ross | hmm | 20:11:19 |
| Alyssa Ross | it shouldn't do | 20:11:22 |
| cole-h | Or maybe it was debugging during development that's been cargo-culted 🤷 | 20:11:23 |
| Alyssa Ross | Specifically broken disallows accessing outPath or drvPath | 20:11:30 |
| Alyssa Ross | As it's implemented with lib.extendDerivation | 20:12:42 |
