| 15 Dec 2025 |
 emily | something like env.REBUILD_HACK = lib.optionalString (stdenv.name == "stdenv-darwin") "sigh";, I think? | 21:00:57 |
| emily | oh: | 21:01:30 |
| emily | pythonSupport ? false,
| 21:01:31 |
| emily | we can probably just condition on that. | 21:01:37 |
| emily | want me to send a PR to check the rebuilds? | 21:01:42 |
 vcunat | Sounds like a good idea. | 21:12:46 |
| emily | https://github.com/NixOS/nixpkgs/pull/471151 | 21:13:04 |
| emily | 774 rebuilds. should be fine. | 21:28:30 |
| emily | vcunat: have you seen more than these couple of incidents lately? | 21:33:42 |
| vcunat | No. | 21:34:00 |
| emily | it's been a very deep abyss to try and figure out what causes those issues so it worries me if it's going to become a common problem | 21:34:30 |
| emily | but I guess it's incentive to dig deeper too :) | 21:34:36 |
| vcunat | But I tend to avoid darwin stuff. Here they were just significant staging-next* regressions which noone looked at. | 21:34:38 |
| vcunat | * But I tend to avoid darwin stuff, so I might've missed them easily. Here they were just significant staging-next* regressions which noone looked at. | 21:35:09 |
| emily | clean-up https://github.com/NixOS/nixpkgs/pull/471160 | 21:44:40 |
 Randy Eckenrode | That’s one that breaks frequently IME. | 22:14:14 |
| Randy Eckenrode | Maybe have a check that xmllint works? | 22:15:19 |
| emily | it seems to only be the Python one this time. | 22:22:31 |
| emily | a default fixup hook to check signing of every Mach-O seems like a good idea if we don't expect to fix the underlying issue soon. | 22:23:01 |
| emily | will try to review your pending PRs today btw | 22:24:24 |
 niklaskorz | pypy on darwin failts to build because it targets macOS 10.13 but uses APIs from 10.15+ | 22:29:29 |
| niklaskorz | why would it target 10.13, I'm confused | 22:29:36 |
| niklaskorz | * pypy on darwin fails to build because it targets macOS 10.13 but uses APIs from 10.15+ | 22:29:48 |
| Randy Eckenrode | It has a dumb default deployment target? | 22:30:00 |
| Randy Eckenrode | Or it conditionally uses those APIs somehow that works on 10.13. | 22:31:15 |
| Randy Eckenrode | https://github.com/pypy/pypy/blob/376cb12babe6bc8f43a550d2981a322513be24de/rpython/translator/platform/darwin.py#L21 | 22:31:54 |
| Randy Eckenrode | Might be a bug since they seem to intend to support 10.13. | 22:32:16 |
| niklaskorz | oh wow | 22:32:17 |
| emily | lol at the very inaccurate comment they keep just appending to | 22:34:12 |
| Randy Eckenrode | Oh, wow. Is 10.15 the real deployment target? | 22:35:07 |
