| 9 Oct 2025 |
 kuflierl | That sounds like a significant issue, maybe keep the old version as default? | 16:25:02 |
 vcunat | * in a several weeks: https://github.com/NixOS/nixpkgs/issues/443568 | 16:25:12 |
| vcunat | I'd say that ship has sailed. There was hope to reduce this a lot during ZHF. | 16:25:39 |
| vcunat | * in several weeks: https://github.com/NixOS/nixpkgs/issues/443568 | 16:25:53 |
 jopejoe1 | We have the ZHF phase of the release to work on fixing those | 16:25:59 |
 leona | Also, we need to update at some point and reverting this and delaying even further won’t help anyone | 16:26:08 |
| vcunat | Well, if I knew how it will go, I would've preferred to delay it by several months. | 16:26:50 |
| vcunat | As I don't see much benefits from having the newer cmake as default. | 16:27:03 |
| vcunat | But at this point I'd say it's way too late to revert. | 16:27:14 |
| vcunat | (and it doesn't seem that bad; lots of work has gone into fixes already) | 16:27:32 |
| kuflierl | In reply to @vcunat:matrix.org
Well, if I knew how it will go, I would've preferred to delay it by several months. Maybe not several months but just after the branch off | 16:27:33 |
| vcunat | That wouldn't ease the amount of work for contributors which was my primary concern. | 16:28:15 |
| vcunat | * That wouldn't ease the amount of work for contributors, which was my primary concern. | 16:28:30 |
| vcunat | Fixes will typically be easy per package, and eligible for backporting, of course. | 16:29:41 |
| kuflierl | In reply to @vcunat:matrix.org
That wouldn't ease the amount of work for contributors which was my primary concern. My concern is more with the stability of the 25.11 release.
Backporting is also more work so I would say it would have been ok | 16:29:41 |
| kuflierl | Making such a major change just before release is bound to make stuff break in untested ways | 16:30:26 |
| vcunat | Still, we have like 6 weeks until branch-off, you know, so I'd hope it will be OK by then 😉 | 16:30:27 |
| kuflierl | In reply to @vcunat:matrix.org
Still, we have like 6 weeks until branch-off, you know, so I'd hope it will be OK by then 😉 Let's hope | 16:30:44 |
| vcunat | Active contributors very often follow master or unstable. | 16:31:23 |
| vcunat | * Active contributors very often follow master or unstable, I believe. | 16:31:51 |
| vcunat | https://github.com/NixOS/nixpkgs/issues/445447 | 16:32:28 |
| kuflierl | In reply to @vcunat:matrix.org
Active contributors very often follow master or unstable. I honestly can not do that. The last time I did that, something significant broke and I didn't really have time to fix stuff on such short notice due to Uni work. I also couldn't downgrade easily due to version incompatibilities.
That's why I like keeping my prod and personal machines on stable.
Though I still prefer unstable for sw development and test VMs | 16:35:41 |
| vcunat | Sure, I meant it as an argument why commonly used stuff should get fixed before the branch-off. | 16:36:41 |
| vcunat | Though yeah, by glancing at that issue, lots of work that doesn't seem worth the benefits 😢 | 16:38:30 |
 dwd | hehe there's 19 more I went to mention on the ticket 👻
surely | 16:40:57 |
| dwd | * hehe there's 19 more I went to mention on the ticket 👻
surely this can be detected | 16:41:12 |
| vcunat | I think some was hacking a script to get all log files and grep them. | 16:41:44 |
| vcunat | (specifically mentioned around that cmake 4 upgrade) | 16:42:10 |
| dwd | aha | 16:42:15 |
| vcunat | given that you can even do nix log -f. attribute... | 16:43:01 |
