| 26 Jun 2025 |
 Alyssa Ross | SGTM | 20:57:02 |
 emily | that seems ideal yes :P | 20:57:04 |
 Grimmauld (any/all) | dangit | 20:57:11 |
| emily | sending the patch upstream will sanity check anyway | 20:57:15 |
| emily | ideally we wait for them to apply it and then switch | 20:57:24 |
| Grimmauld (any/all) | we also do have cmake 4, pkgconf+cmake sounds like a very fun cycle | 20:57:43 |
| Alyssa Ross | We could do CMake first given pkgconf will take time | 20:58:13 |
| emily | CMake and obliterating the bad CMake hook are immediately doable I think | 21:00:05 |
| emily | there's like one blocking issue with the CMake build itself and then it's just jobset tedium | 21:00:29 |
| emily | GCC 15 also probably doable | 21:01:01 |
| emily | LLVM 20 I know Randy is working on as part of his big branch | 21:01:20 |
| emily | btw pkgconf also breaks Meson tests on Darwin and I have no idea why | 21:02:33 |
| emily | we can probably just disable them but it scares me a little | 21:02:43 |
 Tristan Ross | Isn't that just an extension of my PR for bump to LLVM 20 but on the Darwin side of things? | 21:04:13 |
| emily | I think he did it independently beforehand? but I mostly mean I believe he has been fixing breakage along the way since he builds a lot of stuff | 21:05:03 |
| Tristan Ross | Gotcha | 21:05:17 |
| emily | and Darwin consumption of LLVM is changing too wrt libc++ so we probably don't want to run a cycle where we have a libc++ 20 on Darwin that we don't plan to ship | 21:05:48 |
| emily | (though hopefully I will be able to review that and other stuff this weekend anyway) | 21:06:06 |
| Tristan Ross | Great, yeah I believe I got the Linux side of the LLVM 20 bump covered already. | 21:09:05 |
| Tristan Ross | Currently getting back into the GCC NG stuff right now, trying to fix cross compilation. | 21:10:11 |
