| 20 Nov 2022 |
K900 | And the real number of rebuilds is like below 500 | 09:25:54 |
| @omlet:matrix.org left the room. | 09:32:39 |
Artturin | In reply to @k900:0upti.me Something kind of annoying I've been noticing lately: ofborg way overestimates the amount of rebuilds for Qt/KF5 changes like: https://github.com/NixOS/nixpkgs/pull/202019 It doesn't happen with python38Packages packages so look into that https://gist.github.com/GrahamcOfBorg/a062c01a653b883f3c6aa8667b5e8127 | 10:25:19 |
Artturin | The latest 2 python versions are built | 10:25:34 |
K900 | I think it might be not recurseForDerivations | 10:25:51 |
K900 | Maybe we should just do the same to the old Qt versions | 10:26:01 |
| 24 Nov 2022 |
sterni | in principle ofborg should ignore broken stuff unless it evaluates with allowBroken = true since the eval errors get hidden in nix-env | 23:16:14 |
sterni | the passed config should be easy to check in the source, don't know for sure of the top of my head | 23:16:26 |
| 25 Nov 2022 |
| raphi changed their profile picture. | 19:41:52 |
Artturin | cole-h: it's weekend again so a reminder ping https://github.com/NixOS/ofborg/pull/624 | 21:08:16 |
cole-h | That'll probably have to wait until next week -- I've got a decent amount of work to contribute to on my senior project, a midterm, and then finals in a couple weeks. But I'll definitely make a note so I don't forget -- thanks for the reminder. | 21:13:55 |
cole-h | Tentatively planning for next Saturday morning (Pacific) | 21:14:59 |
| 26 Nov 2022 |
| rh changed their display name from rh to ahmed. | 19:19:31 |
| 29 Nov 2022 |
Sandro | Why is this PR which triggers a new eval error for the package showing 0 rebuilds and no error what so ever?
https://github.com/NixOS/nixpkgs/pull/203260 | 16:06:50 |
Artturin | $ nix build ".#python3Packages.modeled"
error: zetup-0.2.64 not supported for interpreter python3.10
(use '--show-trace' to show detailed location information)
| 16:13:10 |
Artturin | perhaps it's caused by the same thing as https://github.com/NixOS/ofborg/issues/610 | 16:13:32 |
Sandro | In reply to @artturin:matrix.org
$ nix build ".#python3Packages.modeled"
error: zetup-0.2.64 not supported for interpreter python3.10
(use '--show-trace' to show detailed location information)
also the input I requested to change is undefined | 16:20:00 |
Sandro | In reply to @artturin:matrix.org perhaps it's caused by the same thing as https://github.com/NixOS/ofborg/issues/610 I also got jinxed by this 2 days ago with an existing package | 16:20:14 |
Sandro | Also python3Packages is still not building. Didn't someone try to fix that already? | 16:21:44 |
Artturin | In reply to @sandro:supersandro.de Also python3Packages is still not building. Didn't someone try to fix that already? yeah me | 16:28:51 |
Artturin | https://github.com/NixOS/nixpkgs/pull/166029 | 16:28:54 |
Artturin | but this introduces more attrs | 16:29:09 |
Artturin | we could also do something like this https://github.com/Artturin/ofborg/commit/5fcea25bc6e213ae2f91dda95b3ad6e813bd1a28 | 16:30:45 |
Sandro | In reply to @artturin:matrix.org but this introduces more attrs so we would boost the rebuild count artificially? | 16:35:30 |
Artturin | yeah | 16:35:41 |
Artturin | better now but still needs refinement https://github.com/NixOS/ofborg/compare/released...Artturin:ofborg:buildpython3packages | 17:36:11 |
Artturin | would be better if the move and .flatten were not there | 17:36:58 |
Artturin | i think | 17:43:51 |
| 30 Nov 2022 |
cole-h | I would be very unhappy to merge that (not that I don't appreciate your work and that this issue exists).
I don't want to hard-code a specific python version that will have to be kept in sync. It also carries the possibility of being incorrect when e.g. targeting a staging after 3.11 has merged into it. Or even a previous release that only has 3.10 but not 3.11 and we bump the alias to 3.11 in Nixpkgs and ofborg.
I don't have any better ideas at the moment, though... | 00:06:48 |
Artturin | Have it as a passthru attribute in python3 and then get it in ofborg perhaps | 01:42:57 |