| 18 Dec 2025 |
|  @adam:robins.wtf left the room. | 12:29:30 |
 K900 | OK reverted xsimd, can anyone else check if thrift builds for them on aa64? | 16:03:30 |
| K900 | It locks up in some test for me | 16:03:41 |
 hexa | https://github.com/NixOS/nixpkgs/pull/469130 | 23:13:47 |
| hexa | A bit sad that we missed the c-ares security update | 23:14:26 |
| 19 Dec 2025 |
 ghpzin | Anyone looked at kdePackages.libkleo ?
https://hydra.nixos.org/build/316811546
(pulled with plasma6 module by default from kde-pim)
Adding gpgme,gpgmepp,libgpg-error to extraPropagatedBuildInputs makes it (and next things) build, but does not feel right.
I assume at least qgpgme should work instead of gpgme | 09:25:18 |
| K900 | This is new and weird | 09:26:36 |
| K900 | Why did this work before | 09:26:39 |
| K900 | Did we touch gpgme this cycle? | 09:26:45 |
| ghpzin | Yes, updated to 2.0:
https://github.com/NixOS/nixpkgs/pull/451898 | 09:27:05 |
| ghpzin | Yes, updated 1.24.3 -> 2.0.1:
https://github.com/NixOS/nixpkgs/pull/451898 | 09:27:26 |
| ghpzin | Yes, updated gpgme: 1.24.3 -> 2.0.1:
https://github.com/NixOS/nixpkgs/pull/451898 | 09:27:45 |
| ghpzin | Yes, updated gpgme: 1.24.3 -> 2.0.1:
https://github.com/NixOS/nixpkgs/pull/451898
Actually that PR changed qgpgme too. | 09:30:48 |
| K900 | It should pull qgpgme via autodeps | 09:35:45 |
| ghpzin | You mean the other way around qgpgme should pull gpgme and other things ?
Or that qgpgme should be in one of the generated kde files ? | 11:05:32 |
| ghpzin | You mean the other way around qgpgme should pull gpgme and other things ?
Or that qgpgme should be in one of the generated kde files ?
Seems like before that PR they were all in one package (qgpgme,gpgme,gpgmepp), so it was not a problem. | 11:09:19 |
| K900 | No, I mean libkleo should pull qgpgme via generated autodeps I think | 12:08:44 |
| K900 | But maybe I'm wrong | 12:08:48 |
| ghpzin | Another weird thing is, libkleo is not the only thing that has qgpgme in dependencies.
But all other things seem to build fine, so are they not using it or just find gpgme in some other way. | 12:12:34 |
| ghpzin | Another weird thing is, libkleo is not the only thing that has qgpgme in dependencies.
But all other things seem to build fine (at least ones that are not blocked on other deps broken on staging-next), so are they not using it or just find gpgme in some other way. | 12:14:04 |
| ghpzin | Another weird thing is, libkleo is not the only thing that has qgpgme in dependencies.
But all other things seem to build fine (at least ones that are not blocked on other deps broken on staging-next).
So are they not using it or just find gpgme in some other way. | 12:14:18 |
| K900 | I think the gpgme setup is pretty cursed | 12:19:51 |
| K900 | Anyway feel free to leave it alone and I'll take a look when I'm less migraine | 13:31:44 |
| ghpzin | Thank you.
If nothing else works as mentioned adding gpgme gpgmepp libgpg-error into extraPropagatedBuildInputs seems to at least build libkleo and all other things depending on it. | 13:34:14 |
| K900 | I don't know why it wants all three | 13:34:59 |
| K900 | Maybe gpgme needs to propagate the other two? | 13:35:06 |
| ghpzin | As I understand all these things (including qgpgme) except for libgpg-error were one and the same before that update PR.
So there was never a question of propagating, it was just a bundle with all cmake .pc files and everything together. | 13:37:06 |
| ghpzin | As I understand all these things (including qgpgme) except for libgpg-error were one and the same before that update PR.
So there was never a question of propagating, it was just a bundle with all cmake, .pc files and everything together.
Not sure how person doing that PR decided on how split should work. | 13:37:44 |
| K900 | @dotlambda thoughts? | 13:38:19 |
| ghpzin | As I understand all these things (including qgpgme that added qt things on top) except for libgpg-error were one and the same before that update PR.
So there was never a question of propagating, it was just a bundle with all cmake, .pc files and everything together.
Not sure how person doing that PR decided on how split should work. | 14:02:46 |
