| 28 Oct 2025 |
 kuflierl | In reply to @ivy:faggot.sh
without having to deal with having a committer go yeah ur the maintainer no shit, but it doesn't even mark me as the maintainer on my prs Thats... Strange | 12:49:19 |
 ivy | then the bot removed it | 12:49:22 |
 K900 | Do you have a link to the PR? | 12:49:30 |
| ivy | https://github.com/NixOS/nixpkgs/pull/456375 i did delete it tho cause i was getting fed up | 12:49:51 |
| ivy | it just kept going no you arent the maintainer, like yes i am | 12:51:41 |
| ivy | if i re open it will it make the bot not attempt to make a pr? | 12:52:41 |
| K900 | @wolfgangwalther:matrix.org any idea why it would do that? | 12:55:07 |
| kuflierl | On a completely different note, i noticed that in my latest PR nix-update seems to update the Go Derivation hash but seems to use the generated hash for the peotobuf Derivation.
https://github.com/NixOS/nixpkgs/pull/456029
How to reproduce:
- check out PR and go back a commit before the update but after the fix.
- run: nix-update -u memos
| 12:55:27 |
| kuflierl | * On a completely different note, i noticed that in my latest PR nix-update seems to update the Go Derivation hash but seems to use the generated hash for the Protobuf Derivation.
https://github.com/NixOS/nixpkgs/pull/456029
How to reproduce:
- check out PR and go back a commit before the update but after the fix.
- run: nix-update -u memos
| 12:55:43 |
| kuflierl | * On a completely different note, i noticed that in my latest PR nix-update seems to update the Go Derivation hash but seems to use the generated hash for the Protobuf Derivation.
https://github.com/NixOS/nixpkgs/pull/456029
How to reproduce:
- check out PR and go back a commit before the update but after the fix.
- run: nix-update -u memos
- observe the diff to my last commit
| 12:57:07 |
| ivy | okay now its saying maintainer!?!?! | 12:58:49 |
| ivy | whattt | 12:58:51 |
| ivy | why didn't it before!?!?!? | 12:59:27 |
| ivy | Wolfgang Walther: thoughts? | 12:59:50 |
 Wolfgang Walther | I looked into this earlier and I expected the label to show up the next time CI would run there. This must have been GitHub returning a wrong (or no) user ID for the PR from their API - all other code looks fine. | 13:01:34 |
| Wolfgang Walther | Also note that the label currently does not matter at all for whether you can merge with the merge bot or not. | 13:01:49 |
| Wolfgang Walther | * Also note that the label does not matter at all for whether you can merge with the merge bot or not. | 13:02:07 |
| Wolfgang Walther | The merge bot looks at different data, and merging would always be possible without the label. The label is just informational for others. | 13:02:28 |
| ivy | i know like obviously i can't merge this because i'm not a committer and the bot didn't make it | 13:04:22 |
| ivy | also for some reason it seems r-ryantm attempts to evaluate the package after updating it, and if its not able to be evaluated on linux it says failed to update leaving it unusable for darwin only programs | 13:07:24 |
| Wolfgang Walther | Well, I interpreted your earlier comment to mean that you thought you wouldn't even be able to merge an automated PR yourself, because the label was missing, too. These two are different, so just because the label was missing now, that would not mean that the merge bot for an automated PR would not work. That's all I'm saying. | 13:08:44 |
| ivy | nah i didnt assume that | 13:09:14 |
| ivy | my thought was that it makes the pr seem less like an easy job for a committer | 13:09:27 |
| ivy | if it doesn't say im the maintainer it incites greater unneeded scrutiny | 13:09:47 |
| ivy | but it doesn't help that nixpkgs-update seems to attempt to evaluate all packages it updates, even if they will not evaluate on that platform | 13:10:45 |
| ivy | see this https://nixpkgs-update-logs.nix-community.org/karabiner-dk/2025-10-13.log | 13:11:07 |
| Wolfgang Walther | I think that's your own update script doing that. You're using common-updater-scripts "update-source-version". Try it's --system= argument to force aarch64-darwin for evaluation. | 13:16:25 |
| Wolfgang Walther | * I think that's your own update script doing that. You're using common-updater-scripts "update-source-version". Try its --system= argument to force aarch64-darwin for evaluation. | 13:16:34 |
| ivy | ahhhh | 13:19:26 |
| ivy | very unclear in documentation of nixpkgs-update that that's how its meant to work | 13:19:45 |
