| 13 Jan 2023 |
 wucke13 | grahamc (he/him): K900 thank you two 😄 | 17:58:30 |
 @grahamc:nixos.org | here are the headers we expect:
proxy_set_header Host $host;
proxy_set_header X-Real-IP $remote_addr;
proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
proxy_set_header X-Forwarded-Proto $scheme;
proxy_set_header X-Request-Base /hydra;
| 17:58:42 |
| @grahamc:nixos.org | so indeed just Host | 17:58:49 |
| @grahamc:nixos.org | sorry for the wildly misleading info :D | 17:58:56 |
| wucke13 | That works like a charm. Thank you for the snippet :) | 18:00:27 |
| @grahamc:nixos.org | https://github.com/NixOS/hydra/blob/master/doc/manual/src/configuration.md#serving-behind-reverse-proxy | 18:01:26 |
| wucke13 | Check. How would the Open GitLab Merge requests input type work in combination with a login-only gitlab instance? Can I use a project access token to make it work? | 18:10:56 |
| 14 Jan 2023 |
|  Nick Cao joined the room. | 13:16:52 |
| 15 Jan 2023 |
 void | any pointers on building home-manager users stuff on hydra? like a small nix file to evaluate ;) | 08:09:35 |
| void | Im recompiling chromium all the time ;( | 08:09:55 |
 shawn8901 | In reply to @void68:matrix.org
Im recompiling chromium all the time ;( Did you override the unwrapped package? If yes, then this is likely the reason you have to compile it. | 13:48:33 |
 Amanda (she/her) | or if you'reon the small channel | 13:48:52 |
| shawn8901 | Yeah true that's another likely reason :) | 13:49:37 |
| shawn8901 | But if your question goes more in the area "how to build my own things on my own hydra instance" if you are on a flake you have to have the thing you want to build in the hydraJobs output | 13:51:42 |
| void | just for expression to eval home.nix in a job. | 22:49:20 |
| void | * just looking for expression to eval home.nix in a job. | 22:49:27 |
| 16 Jan 2023 |
|  vertebralsilence changed their display name from flurie to Alexander Flurie. | 15:14:33 |
| 22 Jan 2023 |
|  Mic92 left the room. | 20:53:57 |
| 24 Jan 2023 |
 vcunat | Minor nit, probably easy to fix by someone who knows hydra's code base:
when showing commit hashes as e.g. "changes" in https://hydra.nixos.org/build/206632301
they're cut too short (6 chars) - and thus often they're not unique anymore in our huge nixpkgs repo. Just 7 would probably be quite good for some time for us. | 09:45:21 |
 das_j | Yes | 09:45:43 |
| das_j | https://github.com/NixOS/hydra/pull/1258 | 09:45:55 |
| vcunat | Oh, and I thought I don't need to search because it's too simple 😄 | 09:47:55 |
| das_j | In reply to @vcunat:matrix.org
Oh, and I thought I don't need to search because it's too simple 😄 $ git rev-parse --short HEAD
d3aa3574382
| 10:16:04 |
| das_j | looks like my git now things 11 characters are appropriate :O | 10:16:19 |
| das_j | * looks like my git now thinks 11 characters are appropriate :O | 10:16:27 |
| das_j | * looks like my git now thinks 11 characters are appropriate for nixpkgs :O | 10:16:40 |
| vcunat | And I thought the suggestion of 12 was excessive. | 10:17:22 |
| vcunat | I expect their calculation is aiming for high probability of all hashes being collision-free. I was OK with with high probability that one hash is collision-free. | 10:25:02 |
| das_j | Hmmm, I think the jobsetinputalts table would be the best place for storing the proper lengths | 10:28:58 |
| das_j | well better yet we should probably use a separate table and insert the length somewhere in hydra-val-jobset around line 820 | 10:36:44 |
