NixOS GNOME | 403 Members | |
| A room for maintainers of GNOME & GNOME-Related desktop environments (xfce, cinnamon, pantheon...) | 93 Servers |
| Sender | Message | Time |
|---|---|---|
| 3 Nov 2023 | ||
| Download tracker-3.6.0.log | 23:39:26 | |
| actually if it builds fine for you it is probably how i am using it in my configuration- i have a member of
| 23:41:02 | |
| anything seem wrong with that? | 23:41:15 | |
| * actually if it builds fine for you it is probably how i am using it in my configuration- i have a member of
| 23:41:36 | |
| 4 Nov 2023 | ||
| That looks fine. I was using that way initially then I had an issue with mismatched GLIBC versions, so I made the gnome branch the primary nixpkgs url (which triggered a massive rebuild) in order to resolve it. I haven't experienced the tracker issue yet but there were some related convos earlier in the chat if you scroll back. The only similar issue was with libsecret, which was caused by 100% CPU usage. I kept restarting the builds and it eventually got there | 00:39:34 | |
| * That looks fine. I was using that way initially then I had an issue with mismatched GLIBC versions, so I made the gnome branch the primary nixpkgs url (which triggered a massive rebuild) in order to resolve it. I haven't experienced the tracker issue yet but there were some related convos earlier in the chat if you scroll back. The only similar issue to yours was with libsecret, which was caused by 100% CPU usage. I kept restarting the builds and it eventually got there | 00:39:43 | |
| alright | 00:48:59 | |
In reply to @donskifarrell:matrix.orgby any chances was the CPU usage error related to a TIMEOUT [...] signal 15 sigterm message? hadn't seen that before this build | 00:52:46 | |
In reply to @shortcutz:matrix.orgAh, just checked and its not quite the same. I was having this issue : https://gitlab.gnome.org/GNOME/libsecret/-/issues/80 | 01:18:45 | |
| thanks anyway | 01:19:48 | |
| not even sure what the issue is, i'm getting different errors every rebuild | 01:20:04 | |
the latest one had to do with a curl failure 🤷 | 01:20:16 | |
shortcutz: the non-deterministic failures sometimes happen more frequently during CPU pressure, you can try building just the affected package if it persists (nix-build https://github.com/NixOS/nixpkgs/archive/gnome.tar.gz -A tracker) | 14:46:00 | |
In reply to @shortcutz:matrix.orgnote that there are also some module changes so you might get issues from that | 14:46:53 | |
| thanks for the tips! | 14:55:37 | |
| i think i can corroborate the cpu pressure thing- i could get much farther (last <100 derivations) built when forcing `max-jobs=1`. maybe that doesnt technically correlate with cpu pressure but it is a fact | 14:56:57 | |
| now my only issue is it running OOM, but i guess i fan fix that by building manually. | 14:57:56 | |
| * now my only issue is it running OOM, but i guess i can fix that by building incrementally. | 15:00:24 | |
| 21:24:41 | ||
| 6 Nov 2023 | ||
In reply to@shortcutz:matrix.orgis this actually the recommended way to do this? wouldn't it make more sense if the gnome branch Nix files are imported using the system pkgs instead of pulling its own pkgs set (which includes a mismatch glibc)? | 05:34:10 | |
| 15:49:19 | ||
| 7 Nov 2023 | ||
| 07:46:25 | ||
| 11:28:40 | ||
In reply to @diamondburned:matrix.orgis there a recommended way? | 17:30:57 | |
| could you explain what you mean by "the gnome branch Nix files are imported using the system pkgs instead of pulling its own pkgs set"? i am afraid im not following | 17:32:34 | |
| 8 Nov 2023 | ||
In reply to @shortcutz:matrix.orgSwitch your whole system to the nixpkgs commit that you want to test. | 06:20:56 | |
In reply to@shortcutz:matrix.orgI just meant using the system nixpkgs to callPackage all the newer packages just so they get built w/ the same stdenv (they need rebuilding anyway) | 06:21:39 | |
| yeah I guess the best way is to rebuild your entire system | 06:21:52 | |
| that will take forever though | 06:21:55 | |
| Well yes. We were waiting for the rebase to make a Hydra jobset... but right now with less than two weeks to 23.11 branchoff, there probably won't even be time to do this kind of pre-testing. | 06:23:02 | |