11 Aug 2024 |
Mic92 | das_j: John Ericson I would like to apply for hydra commit access: https://github.com/nix-community/infra/pull/1349#discussion_r1712899387 | 03:50:17 |
Mic92 | I haven't contributed anything but have been more of a user. However I know some Perl and I would like to integrate nix-eval-jobs into hydra to close the circle | 03:51:43 |
Mic92 | * I haven't contributed anything but have been more of a user. In nix-community we have some smaller deployment that we use for real-world testing. However I know some Perl and I would like to integrate nix-eval-jobs into hydra to close the circle | 03:52:36 |
Mic92 | * I haven't contributed anything to hydra but have been more of a user. In nix-community we have some smaller deployment that we use for real-world testing. However I know some Perl and I would like to integrate nix-eval-jobs into hydra to close the circle | 03:53:00 |
Mic92 | Redacted or Malformed Event | 12:29:31 |
Mic92 | Interesting I cannot see my own message here | 12:30:20 |
13 Aug 2024 |
John Ericson | Mic92: back from travelling, I would support this! :) | 18:42:55 |
19 Aug 2024 |
| @alejandrosame:matrix.org left the room. | 08:51:37 |
das_j | Mic92: invite is coming up :) | 19:40:16 |
das_j | Sorry for the long wait | 19:41:34 |
Mic92 | No problem. I kept myself busy by cleaning up NixOS infra. | 19:51:41 |
21 Aug 2024 |
hacker1024 | Is there a way to tell Hydra not to worry about specific outputs for a (declarative) jobset? E.g. a job's derivation has split out and debug outputs, but I only care about out . If out is available without debug , though, Hydra will still rebuild the derivation, because it tries to realize both output paths. | 08:15:14 |
hacker1024 | * Is there a way to tell Hydra not to worry about specific outputs for a (declarative) jobset? E.g. a job's derivation has split out and debug outputs, but I only care about out . If out is already available without debug , though, Hydra will still rebuild the derivation, because it tries to realize both output paths. | 08:17:04 |
Mindavi | Not that I'm aware of, and how can hydra know out doesn't change? | 11:59:34 |
hacker1024 | In reply to @rick:matrix.ciphernetics.nl Not that I'm aware of, and how can hydra know out doesn't change? Not sure what you mean by that exactly? If out changes, the path will change too. If the path exists, then it does not need to be rebuilt. | 12:06:26 |
22 Aug 2024 |
Mindavi | If debug changes I'm pretty sure out will always change too, unless it's a ca derivation | 08:29:43 |
vcunat | Unless it's a unreproducible change. | 10:07:38 |
vcunat | (same derivation but results not bit-for-bit equal) | 10:08:09 |
25 Aug 2024 |
hacker1024 | Looks like https://github.com/NixOS/hydra/pull/1359 broke Hydra's .narinfo server quite significantly. After fixing two errors with https://github.com/NixOS/hydra/pull/1405, I am now stuck on the following. Does anyone have any ideas?
Caught exception in Hydra::Controller::Root->nar "Undefined subroutine &Hydra::Controller::Root::isValidPath called at /nix/store/kmpplf0rgdh7djzwbb9gbc3mdcc8k6im-hydra-2024-07-09/libexec/hydra/lib/Hydra/Controller/Root.pm line 333."
| 08:12:25 |
Mindavi | Consider adding a test to prevent future regressions at least, but no idea, never used that functionality | 08:41:08 |
27 Aug 2024 |
Mindavi | https://mastodon.delroth.net/@delroth/113029832631860419
FYI: unauthenticated endpoint for triggering evals | 05:03:50 |
| oliver.falvai joined the room. | 10:53:17 |
| @aloisw:kde.org left the room. | 17:26:30 |
das_j | Mic92 and other committers: please temporarily abstain from merging things or pushing to master. we are requesting a CVE ID (again) and it would be lovely if the hashes don't differ from the announcement when we pick to master | 17:51:27 |
Mic92 | Alright | 18:34:12 |
Mic92 | janne.hess: are you working on a fix? | 18:35:40 |
das_j | In reply to @joerg:thalheim.io janne.hess: are you working on a fix? the fact that you are in the infra-build group means you can already see the advisory :D | 18:36:19 |
Mindavi | Lix hydra also has a patch publicly available, I saw | 18:45:21 |
Mindavi | Not sure if it covers everything but at least that endpoint | 18:45:33 |
das_j | CVE ID got assigned and we have published the advisory: https://github.com/NixOS/hydra/security/advisories/GHSA-xv29-v93r-2f5v | 18:57:51 |