| 23 May 2023 |
 Weijia | In reply to @vcunat:matrix.org
Aaah, we can't have 23.05 darwin channel because release critical stuff (wireshark) hasn't been building since January and noone seems to mind 🤦
https://hydra.nixos.org/job/nixpkgs/nixpkgs-23.05-darwin/darwin-tested
https://hydra.nixos.org/job/nixpkgs/trunk/darwin-tested Qt 5 was modified to target SDK 11 earlier this year, while stdenv is at 10.12 | 14:08:02 |
 VladimÃr ÄŒunát | (at least we have ofBorg) | 14:08:04 |
 Ilan Joselevich (Kranzes) | Darwin is a black box and hard to debug on and it gets even worse if you don't have direct access to a Darwin machine. | 14:08:17 |
| Weijia | In reply to @vcunat:matrix.org
Aaah, we can't have 23.05 darwin channel because release critical stuff (wireshark) hasn't been building since January and noone seems to mind 🤦
https://hydra.nixos.org/job/nixpkgs/nixpkgs-23.05-darwin/darwin-tested
https://hydra.nixos.org/job/nixpkgs/trunk/darwin-tested * Qt 5 was modified to target SDK 11 earlier this year, while stdenv is at 10.12 on x86_64-darwin | 14:08:48 |
| Weijia | I suspect almost all Qt 5 packages are broken there for mixing SDKs. | 14:10:16 |
| Weijia | * I suspect almost all Qt 5 packages are broken there for mixing SDKs | 14:10:24 |
| Weijia | Clearly not an issue on aarch64-darwin | 14:11:25 |
| Weijia | But Qt 6 now asks for SDK 12 :-( | 14:11:40 |
 @linus:schreibt.jetzt | In reply to @wegank:matrix.org
Clearly not an issue on aarch64-darwin yet? :> | 14:11:48 |
| VladimÃr ÄŒunát | It has higher SDK in stdenv already. | 14:14:10 |
| Weijia | In reply to @linus:schreibt.jetzt
yet? :> aarch64-darwin only has SDK 11 | 14:14:17 |
| @linus:schreibt.jetzt | right, but surely it'll have a new SDK eventually | 14:14:39 |
| VladimÃr ÄŒunát | * It has higher SDK in stdenv already (higher than x86). | 14:14:40 |
 raitobezarius | Good, all my servers are running 23.05 now :) | 14:16:13 |
| Weijia | In reply to @linus:schreibt.jetzt
right, but surely it'll have a new SDK eventually Yeah, a lot: https://github.com/NixOS/nixpkgs/pull/229210 | 14:17:08 |
| Weijia | But I'm not sure if we can really move Qt to SDK 12 on aarch64-darwin without a global SDK bump | 14:18:58 |
| raitobezarius | https://github.com/NixOS/nixpkgs/pull/233625 | 14:47:46 |
| raitobezarius | hexa: ^ isn't this too late? :/ | 14:47:51 |
 hexa | tbh: we have to accept whatever allows operators to continue their deployment | 14:48:30 |
| raitobezarius | hm not sure what this means here, I don't know enough operators in k8s | 14:49:11 |
| raitobezarius | * hm not sure what this means here, I don't know enough operators' version constraints in k8s | 14:49:17 |
| VladimÃr ÄŒunát | Maybe for the following releases we should explicitly clarify whether removing a package is a "breaking change". | 14:51:34 |
| raitobezarius | Yeah, IMHO, I'd rather avoid removing any package in the beta phase | 14:52:01 |
| VladimÃr ÄŒunát | * Maybe for the following releases we should explicitly clarify whether removing a package (which built so far) is a "breaking change". | 14:52:10 |
| raitobezarius | People are starting to upgrade to 23.05-beta, feeling it's not hard to fix, to discover commits later that the final release is incompatible | 14:52:28 |
| raitobezarius | And I feel like this is a non-starter in terms of UX | 14:52:35 |
| raitobezarius | (LLVM uses to do this kind of stuff in the release candidates and I hate it) | 14:52:48 |
| raitobezarius | * (LLVM do this kind of stuff in the release candidates and I hate it) | 14:52:55 |
| VladimÃr ÄŒunát | In reply to @vcunat:matrix.org
Maybe for the following releases we should explicitly clarify whether removing a package (which built so far) is a "breaking change". I'd probably think that it is a breaking change. Or at least when the alternatives aren't fully compatible. | 14:54:10 |
| Ilan Joselevich (Kranzes) | Can I merge and backport this? https://github.com/NixOS/nixpkgs/pull/232459 | 14:54:52 |
