| 28 Mar 2023 |
 raitobezarius | awesomeeeeeee | 14:00:51 |
 baloo | does that still run once a day? | 17:19:08 |
| baloo | or is that still the same infrastructure behind, just a different frontend? (in any case, thank you!) | 17:19:41 |
 raboof | the "once a day" infra hasn't run since June, this now runs "when I feel like it" 😃. the goal is to run it on shared infra (possibly nix-community), until then I'm also happy to take PR's to https://github.com/NixOS/reproducible.nixos.org | 18:27:44 |
| 29 Mar 2023 |
| raboof | I'd like to experiment with using GitHub Projects for tracking the issues that are relevant for each milestone: https://github.com/orgs/NixOS/projects/30/views/1 - LMK if anything seems wrong/incomplete! (long-term it would be great to link this from the reports themselves somehow) | 10:31:31 |
| 3 Apr 2023 |
|  @kranzes:matrix.org joined the room. | 01:20:01 |
|  docRekd joined the room. | 17:09:42 |
|  @peddie:matrix.org joined the room. | 17:35:14 |
|  @ulli:hrnz.li joined the room. | 17:57:55 |
|  javbit joined the room. | 21:14:00 |
| 4 Apr 2023 |
 Artturin | https://github.com/NixOS/nixpkgs/pull/224670 | 15:35:41 |
| 5 Apr 2023 |
|  tea joined the room. | 08:08:42 |
| tea | My derivation sometimes fails to build because an unknown factor causes a hash change.
If I build with --keep-failed, it does keep the directory of the build, but not the thing (a targz) in $out, which is the thing that fails. | 08:11:04 |
| tea | * My derivation sometimes fails to build because an unknown factor causes a hash change.
If I build with --keep-failed, it does keep the directory of the build, but not the thing (a targz) in $out, which is the thing that has a mismatching hash.
The problem is that the hash changes very frequently and sometimes I cannot get there fast enough to change to the current "correct" hash
| 08:11:44 |
| tea | Is there a way to get $out ignoring the hash mismatch? | 08:11:54 |
|  ribosomerocker joined the room. | 08:14:20 |
| raboof | in 'derivation X may not be deterministic: output Y differs from Z', Y contains the $out of the 'previous' build and Z contains the $out of the 'fresh' build. Is it possible to share the problematic derivation? | 08:14:51 |
| tea | But I cannot get it to build correctly, because the hash seems to change like every second | 08:15:49 |
| tea | it is a FOD | 08:18:02 |
| raboof | that's the hash of something that is used in the derivation, I guess, right? like in fetchurl or something? or in some other fixed-output derivation? | 08:18:04 |
| raboof | it sounds like the FOD is not the derivation you're nix-building, but a dependency of the derivation you're nix-building. If you nix show-derivation, you might be able to spot the FOD in the inputDrvs and build it independently? | 08:20:09 |
| tea | I am building it independently. | 08:20:45 |
| tea | the full story is i am a flutter maintainer and the new flutter version broke our flutter.mkFlutterApp automation.
mkFlutterApp has a deps FOD, to fetch app deps using its own package manager, which is now not reproducible anymore.
The hash of any deps derivation is now changing very rapidly and I'd want to fix it. | 08:23:01 |
| raboof | hmm, gotcha, I didn't realize FOD's fail so early, that indeed makes it tricky | 08:26:22 |
|  MGlolenstine joined the room. | 08:27:18 |
| tea | I guess for now i am going to make the tarball not output to $out but a local file | 08:27:36 |
| tea | which is kept by --keep-failed | 08:27:40 |
| raboof | yeah or printing find $out | xargs md5sum in the postFixup might put enough useful information in the build logs | 08:29:17 |
| raboof | or find /path/to/tarball/inputs :) | 08:30:29 |
|  @rimuru:gentoo.chat changed their profile picture. | 12:30:20 |
