| 21 May 2026 |
 vcunat | I didn't know that staging-nixos was expected. | 08:04:56 |
 K900 | Uhh | 08:05:26 |
| K900 | I merged the systemd fix into staging-nixos | 08:05:30 |
| K900 | And then staging-nixos into master | 08:05:39 |
| K900 | The only other change on staging-nixos is a ty upgrade | 08:05:48 |
| K900 | So I don't think we're going to have a problem there | 08:05:57 |
| vcunat | Ah, I overlooked the target branch. | 08:06:28 |
 Emil Thorsøe | In file included from /nix/store/q9ksx8c79jfj1cawwwzmpq3qapvpvab7-glib-2.88.1-dev/include/glib-2.0/glib/gstring.h:37,
from /nix/store/q9ksx8c79jfj1cawwwzmpq3qapvpvab7-glib-2.88.1-dev/include/glib-2.0/glib/giochannel.h:36,
from /nix/store/q9ksx8c79jfj1cawwwzmpq3qapvpvab7-glib-2.88.1-dev/include/glib-2.0/glib.h:56,
from ../src/vtegtk.cc:44:
../src/vtegtk.cc: In function 'char* vte_terminal_get_selection(VteTerminal*)':
../src/vtegtk.cc:4362:59: error: 'to_integral' is not a member of 'vte'
4362 | return g_strdup (IMPL(terminal)->m_selection[vte::to_integral(vte::platform::ClipboardType::PRIMARY)]->str);
| ^~~~~~~~~~~
/nix/store/q9ksx8c79jfj1cawwwzmpq3qapvpvab7-glib-2.88.1-dev/include/glib-2.0/glib/gstrfuncs.h:324:38: note: in definition of macro 'g_strdup'
324 | #define g_strdup(x) g_strdup_inline (x)
Does someone recognize off-hand, why staging-next merge invoked this error, I don't know much C++, but somehow looks like internal inconsistency, that I would not expect to compile anywhere | 10:48:53 |
| vcunat | Perhaps write at least which package is triggering this glib error? | 10:51:41 |
| Emil Thorsøe | vte | 10:51:58 |
| vcunat | There was a major glib upgrade, so that's the trigger, most likely. | 10:52:00 |
| Emil Thorsøe | can it affect this, if it's C library? | 10:52:36 |
| Emil Thorsøe | but I guess | 10:52:42 |
| vcunat | Well we do see that they use this C header from this .cc file (transitively?) | 10:53:44 |
| vcunat | * Well we do see that they use this C header from this .cc file (transitively maybe) | 10:53:48 |
| vcunat | But on hydra all vte seems green: https://hydra.nixos.org/eval/1825558?filter=vte | 10:54:01 |
| K900 | Yes? glib is a C library | 10:54:02 |
| Emil Thorsøe | interesting | 10:54:34 |
| vcunat | C++ and C interop nicely. | 10:54:46 |
| Emil Thorsøe | wonder if I have some magic vte of my own making, I have no idea about why I would have that | 10:54:58 |
| Emil Thorsøe | I'll check why-depends | 10:55:09 |
| vcunat | * C++ and C interop nicely. (but that's tangential here anyway) | 10:55:27 |
 ghpzin | From code looks like:
https://github.com/NixOS/nixpkgs/blob/fea3d623979a5545017061bdfc828faa441806b1/pkgs/applications/terminal-emulators/termite/default.nix#L21 | 10:56:19 |
| vcunat | It's a shame that we don't have something like why-depends which prints attribute names or points into interesting places in .nix files. | 10:56:24 |
| ghpzin | From code looks like vte-ng pulled by termite:
https://github.com/NixOS/nixpkgs/blob/fea3d623979a5545017061bdfc828faa441806b1/pkgs/applications/terminal-emulators/termite/default.nix#L21 | 10:57:08 |
| vcunat | vte-ng = vte.overrideAttrs (attrs: {
patches = attrs.patches or [ ] ++ [
| 10:58:02 |
| vcunat | so that vte variant doesn't have its own hydra job | 10:58:15 |
| Emil Thorsøe | termite is my guess | 10:58:17 |
| Emil Thorsøe | verifying | 10:58:20 |
| vcunat | * vte-ng = vte.overrideAttrs (attrs: {
patches = attrs.patches or [ ] ++ [
(from termite expression)
| 10:58:45 |
