| 18 Nov 2025 |
 Winter | this is on aarch64-darwin native | 03:03:24 |
| Winter | and above that, wiggles’ comment | 03:03:31 |
| Winter | i definitely didn’t make it clear enough (sorry!), but wiggles did | 03:05:11 |
 Tristan Ross | Oh, it didn't load for me. Had to click the "load more" | 03:05:43 |
| Tristan Ross | 
Download image.png | 03:08:16 |
| Tristan Ross |
channel blockers: aarch64-darwin: some
| 03:08:22 |
| Tristan Ross | *
channel blockers: arm64-apple-darwin: some
| 03:08:32 |
| Tristan Ross | Hmm, might be a good idea to get this fixed | 03:08:45 |
| Tristan Ross | Let's see what the release managers think first though | 03:09:09 |
| Winter | realistically we could wait a cycle but i’d rather not. | 03:12:44 |
| Winter | it’s also kind of technically probably not a darwin specific issue | 03:12:57 |
| Winter | but i haven’t looked into it too much | 03:13:03 |
 dish [Fox/It/She] | its not, https://github.com/nixos/nixpkgs/issues/461884#issuecomment-3543238002 mentions cross-compiling for aarch64-linux on x86_64-linux | 03:14:59 |
| dish [Fox/It/She] | its definitely not, https://github.com/nixos/nixpkgs/issues/461884#issuecomment-3543238002 mentions cross-compiling for aarch64-linux on x86_64-linux | 03:15:11 |
| dish [Fox/It/She] | its definitely not, https://github.com/nixos/nixpkgs/issues/461884#issuecomment-3543238002 mentions cross-compiling for aarch64-linux on x86_64-linux and encountering this bug | 03:15:20 |
| Winter | i was under the assumption that that was a separate but similar bug… guess not, didn’t really read too much into it after discovering the cursedness | 03:18:10 |
|  Uraraka ~ Ochaco changed their display name from Peri to Uraraka ~ Ochaco. | 05:23:33 |
 vcunat | Python is a stdenv rebuild on darwin, so fixing that would most likely imply delaying the release, as it's about the same speed as building completely everything from scratch on all systems. | 06:36:44 |
 leona | I haven't seen that this also caused linux fallout, I read the issue only after the first comment and haven't subscribed apparently. I'm not sure what to do. Could we try to leave the python used for darwin bootstrap to the broken one and "only" rebuild anything after that? That would be 144k rebuilds, so at least somewhat better | 08:15:08 |
| leona | I don't like changing the release date once more to (then probably) December | 08:15:43 |
| leona | I'd like to know jopejoe1s opinion tho.. | 08:15:46 |
 jopejoe1 (4094@39c3) | Would prefre not rebuilding stdenv this late into the last staging-next cycle before branch off, and would want to have the rebuild count as small as possible, so going with somethig like @leona:leona.is suggested and then doing a revert and full fix on staging (with backport). | 08:34:00 |
| leona | i can try something out in around 3 hours, if someone beats me to that that's very welcome. I don't know very much about darwin stdenv at the moment | 08:36:24 |
| leona | from what I read, this should be possible, but I'm not sure how much work is needed | 08:36:47 |
| vcunat | Side not: IMO it's wrong to leave big fixes to release managers. | 08:59:13 |
| vcunat | * Side not: IMO it's (generally) wrong to leave big fixes to release managers. | 08:59:35 |
| leona | nargh | 13:05:32 |
| leona | at least currently I don't get it to work | 13:05:44 |
| leona | and my circulation is a bit off today, so I'm not at my best either. | 13:06:35 |
| leona | the idea is relatively simple, having a seperate python3Stdenv package that is used for llvm and other things in the darwin stdenv. but for now it's still at various kinds of broken | 13:07:28 |
