| 21 May 2022 |
Vladimír Čunát | Ah right. I have the "Infra Team" term associated with a much narrower set of people (and corresponding private room). | 20:26:22 |
Janne Heß | In reply to @vcunat:matrix.org Ah right. I have the "Infra Team" term associated with a much narrower set of people (and corresponding private room). Ah, I don't think I can expect every release manager to be part of that private room ;) | 20:27:14 |
Vladimír Čunát | The room you link is much better for this. | 20:27:48 |
Vladimír Čunát | I'm not on the infra team itself, but I can create jobsets (in nixpkgs and nixos projects). Though I'm not sure how many other such people exist. | 20:28:32 |
Janne Heß | Probably not too many | 20:29:11 |
Vladimír Čunát | Release managers commonly do get restart/cancel privileges at least, right? | 20:31:31 |
Janne Heß | In reply to @vcunat:matrix.org Release managers commonly do get restart/cancel privileges at least, right? Yeah, I have everything except admin and create-projects | 20:31:53 |
Janne Heß | So also the bump privilege | 20:32:02 |
Vladimír Čunát | Right, that's relevant. | 20:32:48 |
Janne Heß | Vladimír Čunát: about your comment: is there any reason to lower the scheduling shares rather than just disabling the jobset altogether? | 20:47:31 |
Vladimír Čunát | People still push to them sometimes. | 20:48:19 |
Janne Heß | ummm what? | 20:48:26 |
Vladimír Čunát | Say
commit + 1a7c2f41c3bf6d7fd443e2453e7b840cdc413375 (h/release-21.05, h/nixpkgs-21.05-darwin, h/nixos-21.05-aarch64)
Author: Robert Hensing <robert@roberthensing.nl>
AuthorDate: Thu Apr 7 11:05:28 2022 +0200
Commit: Robert Hensing <robert@roberthensing.nl>
CommitDate: Mon May 2 09:51:01 2022 +0200
pkgs: Add _type = "pkgs"
(cherry picked from commit ad1e2500efd0aa49b0dc7427bf69d4879f3b0ff5)
(cherry picked from commit aec730a0af4c977513ce28236cbecaca72af6901)
| 20:49:04 |
Vladimír Čunát | That's to 21.05 just several weeks ago. | 20:49:21 |
Vladimír Čunát | * That's to 21.05 just three weeks ago. | 20:49:39 |
Vladimír Čunát | To me, removal of support is just removal of commitment/guarantees about maintenance. | 20:50:18 |
Vladimír Čunát | But if someone still wants to do it, I can't see why block it. But naturally it should have lower priority on the shared infra. | 20:50:55 |
Vladimír Čunát | * But if someone still wants to do it, I can't see why block it. Naturally it should have lower priority on the shared infra. | 20:51:05 |
Janne Heß | I've had this discussion in another context, but I think changing EOL releases makes it seem a lot like it's still (at least partially) maintained | 20:51:49 |
Vladimír Čunát | If they are, is it bad that they look like it? | 20:52:39 |
Vladimír Čunát | Maybe if it's confusing, yes. | 20:53:11 |
Janne Heß | Well if they were, we should have a consistent wording. status.nixos.org clearly states these as EOL | 20:53:15 |
Janne Heß | yeah | 20:53:16 |
Vladimír Čunát | Release notes stress the dates visibly, more importantly. | 20:53:39 |
Vladimír Čunát | But I can't estimate how well this information gets passed in practice. Our point of view is too biased to see ourselves. | 20:54:29 |
Janne Heß | Hm, if it was me, I'd disable the jobsets (maybe leave the previous release?) and protect the release- branch. This way we can get around all possible accidents. But that's my opinion and I know it's the strictest one and there are people who would probably disagree | 20:54:56 |
Janne Heß | * Hm, if it was me, I'd disable the jobsets (maybe leave the previous release?) and protect the release- branch. This way we can get around all possible accidents. But that's my opinion and I know it's the strictest one and there are people who would disagree | 20:55:07 |
Vladimír Čunát | Right, e.g. for older than 21.05 right now that makes sense to me. | 20:57:25 |
Vladimír Čunát | Though very little impact I guess, as I'd hope noone uses such old channels/branches anymore. | 20:58:25 |
Vladimír Čunát | * Though very little impact I guess, as I'd hope noone follows such old channels/branches anymore. | 20:58:34 |