| 21 May 2026 |
hexa | .so cannot work | 13:37:23 |
hexa | referencend from a dylib | 13:38:11 |
hexa | great stuff | 13:38:13 |
Randy Eckenrode | .so is fine. The install name of libhdf is wrong. It needs to be an absolute path not an rpath. | 13:58:51 |
| 22 May 2026 |
| Unhooked Spirits joined the room. | 05:57:13 |
| Jon Hermansen changed their display name from jonhermansen to Jon Hermansen. | 19:18:45 |
hexa | https://hydra.nixos.org/job/nixpkgs/unstable/firefox-esr-140-unwrapped.aarch64-darwin | 21:08:26 |
hexa | Ihar Hrachyshka 140.11.0esr startet failing after the staging-next merge | 21:08:42 |
hexa | on both darwins | 21:08:58 |
hexa | https://cache.nixos.org/log/mfg292hnvkh0zd16h7m5xm2bjwwh52zj-firefox-unwrapped-140.11.0esr.drv | 21:10:08 |
hexa |
error: could not compile gecko-profiler (lib) due to 1 previous error
| 21:10:12 |
K900 | I bet it's the rustc update | 21:10:49 |
hexa | yeah, with certainty | 21:11:12 |
hexa | 1.94.x -> 1.95.0 | 21:11:19 |
hexa | I have no capacity to take care of that | 21:11:33 |
hexa | https://paste.lossy.network/WSLLAOC4XUVLMMIS2R7L4SSILY | 21:26:20 |
hexa | https://bugzilla.mozilla.org/show_bug.cgi?id=2041825 | 21:31:06 |
hexa | filed | 21:31:09 |
ghpzin | Looks a lot like:
https://bugzilla.mozilla.org/show_bug.cgi?id=2038918
except it is a different type and platform | 21:38:32 |
| 23 May 2026 |
Randy Eckenrode | Stage is open for breaking changes again, right? What’s our timeline on dropping x86_64-darwin support? | 00:41:13 |
Randy Eckenrode | * | 00:41:22 |
| @not-jack:matrix.org left the room. | 09:17:46 |
hexa | emily wanted to be pinged 3 days before the new staging-next opens up iirfc | 12:08:14 |
hexa | Redacted or Malformed Event | 12:08:17 |
hexa | Redacted or Malformed Event | 12:08:21 |
| @lrsauvette:sauvette.de left the room. | 15:40:54 |
| @bryan.bennett:matrix.org left the room. | 18:23:15 |
| 24 May 2026 |
emily | I have the core changes all ready to go, yeah (and open PRs already). the WIP stuff I wanted a few days notice for is just finishing triaging stuff that dropping a platform regresses (e.g. a lot of binary packages have update scripts that will break when x86_64-darwin does) – I have commits for it but they're messy. | 18:01:10 |
emily | I did want us to resolve the Wine question though. carving out just that package and its revdep tree would be easy / minimal maintenance and builder load, especially if we drop it as soon as it poses issues or we have a viable replacement and if some of the Linux-centric deps can be dropped. it's just slightly awkward timing – I suspect that CrossOver is going to come out with something during the macOS 27 lifetime given their collaboration with Apple and the Rosetta changes coming in macOS 28. having a period between now and then where macOS users can't get an up-to-date / supported Wine at all feels a bit unfortunate.
it's just a matter of whether I drop the few changes that would need to be elided to carve out a pkgs.wine specifically a la Mesa while keeping everything else behind the error/not-being-built gate. (I could see if we can do it via cross too, which would cut out even more of the support/builds.)
| 18:05:50 |
emily | I was really hoping that we'd have a working native Wine + FEX/Rosetta situation by now tbh | 18:07:05 |