| 16 Jun 2025 |
emily | what changed about the setup to start blowing up? did stuff start looking for QDoc suddenly? | 09:53:52 |
K900 | Because again, those Qt packages aren't found separately | 09:53:55 |
K900 | There's a big FindQt.cmake helper | 09:54:03 |
K900 | That takes a list of modules | 09:54:07 |
K900 | And finds all of them or none of them | 09:54:11 |
K900 | And somehow adding qttools breaks it | 09:54:19 |
K900 | The reason I added qttools is because a lot of frameworks have proper API docs now | 09:54:38 |
K900 | And Gear and Plasma will as well | 09:54:44 |
K900 | In the next release | 09:54:47 |
K900 | So the builds start requiring qdoc | 09:54:55 |
emily | I see | 09:55:33 |
emily | does ECM ship its own FindQt machinery or are we talking about the upstream one? | 09:56:00 |
K900 | Upstream | 09:56:05 |
emily | hmm. I believe CMake has arguments to debug the package finding process. unfortunately I forget what they are | 10:02:22 |
K900 | Well I think it at least works now | 10:20:28 |
K900 | Trying to build my flake with it | 10:21:00 |
K900 | Once that's done I'll deploy it to a machine and then start nixpkgs-review on the big computer | 10:21:12 |
K900 | OK I think I'm giving up on moving deps and I'll just add them | 12:43:15 |
K900 | And we can revert the whole commit once if we ever figure this out properly | 12:43:31 |
K900 | Note to self (or anyone willing to snipe): we should probably fix drkonqi to look up unwrapped executable names | 12:50:59 |
Sandro 🐧 | I've pushed the combined PR commits onto my fork. Probably going to build the system later. | 13:51:25 |
K900 | I'm building everything on that branch now | 14:03:20 |
K900 | God it sucks | 14:03:29 |
K900 | So many rebuilds | 14:03:33 |
K900 | All because cmake is ass | 14:03:47 |
K900 | Can we have meson for qt7 | 14:04:42 |
K900 | Pretty please | 14:05:05 |
Marie | oh yes please | 14:05:06 |
K900 | I am nearly done with what I hope will be the last everything rebuild though | 14:06:38 |
K900 | Oooooor not | 14:20:59 |