| 5 Sep 2024 |
 Moritz Sanft | Just to confirm for a talk: At some point, the minimal ISO was reproducible, right? | 14:11:57 |
 raboof | yes (https://discourse.nixos.org/t/nixos-reproducible-builds-minimal-installation-iso-successfully-independently-rebuilt/34756) - AFAIK it still is, but I'll admit I haven't tried since June | 14:13:23 |
 emily | I thought jfsutils was scuppering it | 14:13:42 |
| raboof | yes and no: we have a strong indication that jfsutils sometimes produces a different output (https://github.com/NixOS/nixpkgs/issues/276433), but it's rare enough that I have never actually ran into that when trying to reproduce the ISO | 14:17:06 |
| raboof | so we're tracking it and want to fix it to avoid false negatives, but it didn't prevent us from reproducing the ISO so far | 14:18:09 |
| raboof | (aka worst-case the issue would cause us not to trust a build that was actually safe, it will never cause us to trust a build that was actually unsafe) | 14:21:03 |
| emily | maybe I should get around to removing JFS to help you out :) | 14:21:15 |
| emily | https://github.com/NixOS/nixpkgs/pull/339821 | 15:01:31 |
| emily | merry christmas | 15:01:33 |
 Philip Taron (UTC-8) | I merged it. | 18:30:22 |
| emily | I've noticed lately that the key to getting things merged in Nixpkgs is to have a PR message of approximately 5 times the length of the diff | 18:31:41 |
| Philip Taron (UTC-8) | It does help that it reads like a motion made with careful forethought and background. In this case, it was actually an anime gif from the release manager that sealed the deal. | 18:33:51 |
 Kamilla 'ova | In reply to @raboof:matrix.org
yes and no: we have a strong indication that jfsutils sometimes produces a different output (https://github.com/NixOS/nixpkgs/issues/276433), but it's rare enough that I have never actually ran into that when trying to reproduce the ISO by the way, this problem is most likely caused by the gnu&llvm strip bug, since I can't reproduce this very strange output with dontStrip = true (after 1000+ rebuilds, with dontStrip = false it fails to reproduce under ~300 rebuilds)
and this is the end, my strength was not enough for further debugging :(
| 20:04:39 |
