| 29 Aug 2023 |
 hexa | pushed the remaining releases and running the tests now on aarch64-linux and x86_64linux | 13:48:35 |
| hexa | * pushed the remaining releases and running the tests now on aarch64-linux and x86_64-linux | 13:48:38 |
| hexa | https://www.mozilla.org/en-US/firefox/117.0/releasenotes/ | 13:49:31 |
 vcunat | Release notes already! | 13:49:45 |
| hexa | yeah, it's tuesday already 🙂 | 13:49:55 |
| hexa | but still early, agreed | 13:50:00 |
| hexa | https://github.com/NixOS/nixpkgs/pull/252170 | 16:44:20 |
| 30 Aug 2023 |
| vcunat | https://github.com/NixOS/nixpkgs/pull/252274 | 06:18:50 |
| 3 Sep 2023 |
| hexa | the firefox tests on trunk-combined are failing with too many open files | 01:09:04 |
|  Emily joined the room. | 01:10:03 |
| hexa | can't reproduce locally | 01:11:08 |
| hexa | but this blocks the channel | 01:11:11 |
| hexa | restarted three times so far | 01:11:20 |
| hexa | 
Download image.png | 08:38:55 |
| hexa | this is going great | 08:38:56 |
| vcunat | It's certainly worse than usual. | 08:42:42 |
| vcunat | (I normally do a few firefox test restarts per eval) | 08:43:00 |
| vcunat | Ah, I see. | 08:44:04 |
| vcunat | In recent past it looked like failing unless built on t4b machine (which is set up quite differently). | 08:45:31 |
| vcunat | But the machine is unreachable right now. | 08:45:44 |
| vcunat | (Which I did know. I planned to look into it physically today anyway.) | 08:46:16 |
| vcunat | It would be best if the test wasn't so sensitive. | 08:47:02 |
| vcunat | * It would be best if the test wasn't so sensitive, of course. | 08:47:05 |
| hexa | can we lift the fd constraint somehow in nixos tests? | 08:56:12 |
| hexa | because this is really weird | 08:56:20 |
| hexa | why is it sufficient on my builders, but not on hydra's when both essentially start a naked nixos in qemu | 08:56:42 |
| vcunat | (I have no idea how this happens.) | 09:01:02 |
| hexa | OK, will dig into it a bit once i can sit down | 09:19:00 |
| vcunat | For now unblocked. Fixed t4b and got lucky scheduling it there on the first retry. But it will be nice if you address the fragility anyway, as restarts are needed often. | 11:54:57 |
| hexa | for me locally the resource limit for nofile on the test script is 1048576 | 20:32:46 |
