| 21 May 2026 |
 Randy Eckenrode | I wonder if I should try reusing the existing Swift 5.10.1 support to bootstrap Swift 6.x. It would get rid of the problem of depending on the LLVM derivations. The upstream binaries are extremely annoying to work with in Nixpkgs. | 20:01:32 |
| Randy Eckenrode | (Speaking of source bootstrap.) | 20:01:42 |
| 22 May 2026 |
 llakala | I'd appreciate feedback on https://github.com/NixOS/nixpkgs/issues/522442 | 14:55:14 |
| llakala | this will give us some free performance on every mkDerivation call | 14:56:05 |
|  Alex Epelde joined the room. | 18:52:10 |
|  Jon Hermansen changed their display name from jonhermansen to Jon Hermansen. | 19:18:33 |
| 23 May 2026 |
|  mtmn joined the room. | 15:22:45 |
| 5 Jun 2026 |
|  9lore changed their display name from 9lore to 9lore | call LORE. | 14:34:57 |
|  zimward changed their display name from zimward to zimward @GPN24. | 19:37:27 |
| 6 Jun 2026 |
|  WorldsGoneMad joined the room. | 19:53:27 |
| 7 Jun 2026 |
| zimward changed their display name from zimward @GPN24 to zimward. | 21:07:16 |
| 8 Jun 2026 |
| 9lore changed their display name from 9lore | call LORE to 9lore. | 00:35:22 |
| 11 Jun 2026 |
 whispers [& it/fae] | out of curiosity, is gcc 16 in the picture for default for 26.11, or would we expect that to be a 27.05+ thing? | 15:45:02 |
 K900 | Probably will be in 26.11 | 15:45:50 |
| K900 | Unless it's a huge unusable mess | 15:45:57 |
| 12 Jun 2026 |
|  Sapii/Saperson changed their display name from Sapii to Sapii/Saperson. | 01:24:23 |
| 13 Jun 2026 |
|  klea (she/her) joined the room. | 21:43:54 |
| 15 Jun 2026 |
|  Eleina joined the room. | 06:11:13 |
| 18 Jun 2026 |
| whispers [& it/fae] changed their display name from whispers [& it/fae] to meow meow. | 18:45:58 |
| whispers [& it/fae] changed their display name from meow meow to whispers [& it/fae]. | 19:11:45 |
| 21 Jun 2026 |
| llakala | followup to the issue posted a few weeks ago - https://github.com/NixOS/nixpkgs/pull/533770 is now good for review | 02:15:40 |
| llakala | gives a really great perf diff | 02:15:50 |
| 22 Jun 2026 |
|  flumffy joined the room. | 02:55:26 |
| flumffy | o/ | 03:00:21 |
| flumffy | I've talked to John Ericson about getting the OpenBSD fixes packaged for a while now, but haven't had time to clean it up enough for nixpkgs (there's some ugly IFD and patches to the llvmPackages components). | 03:02:01 |
| flumffy | I'm not sure whether importing all of the upstream LLVM patches wholesale is the best approach; it does give us a compiler that's as close to what you would get on a native OpenBSD machine at the cost of some miscellaneous patches that are probably unneeded. | 03:04:25 |
| flumffy | It also introduces a need to pass a __FakeBSD__ define into the LLVM build since some of the patches which are required for binaries that will execute properly have their changes gated with an __OpenBSD__ define, while others don't. | 03:05:00 |
 Alyssa Ross | For other BSDs we have not patched compilers. | 07:52:48 |
| Alyssa Ross | Why do these patches need to be in Nixpkgs at all, rather than in LLVM upstream? | 07:52:58 |
| flumffy | The OpenBSD developers maintain a set of patches to the compiler for security reasons. I'm not sure if they've made any attempts at upstreaming said patches, nor am I sure they ever plan to. | 07:55:19 |
