| 2 Dec 2022 |
 cbwang | Perhaps some day, some one would publish this: he compiled an entire NixOS from source code, except the 256-bytes stage0 is binary. Then he publishes his (gpg-signed) entire chain of trust, how he bootstrapped stage0 -> stage1 -> gcc -> nix -> ... NixOS iso, and telling the world that anyone can reproducible this chain of trust. | 18:33:52 |
| cbwang | * Perhaps some day, some one would publish this: he compiled an entire NixOS from source code, except the 256-bytes stage0 is binary. Then he publishes his (gpg-signed) entire chain of trust, how he bootstrapped stage0 -> stage1 -> gcc -> nix -> ... NixOS iso, and telling the world that anyone can reproduce this chain of trust. | 18:40:42 |
 Foxboron | In reply to @cbwang:matrix.org
Besides, if we really can achieve that, then we are going to have the first host OS that all the binaries are free Guix is already doing this fwiw | 18:50:28 |
| Foxboron | They started https://bootstrappable.org/ as a subproject under reproducible builds a few years ago. They have also been working hard on reducing the compiler stages beyond the seed binary with work on GNU mes C | 18:51:51 |
| cbwang | In reply to @foxboron:archlinux.org
Guix is already doing this fwiw However from https://data.guix.gnu.org/repository/1/branch/master/latest-processed-revision/package-reproducibility it seems that more than 10% of Guix packages are not reproducible | 18:54:03 |
 atemu12 | cbwang: I wouldn't be surprised if was a similar number for us aswell. | 18:57:58 |
| Foxboron | In reply to @cbwang:matrix.org
However from https://data.guix.gnu.org/repository/1/branch/master/latest-processed-revision/package-reproducibility it seems that more than 10% of Guix packages are not reproducible Making the seed bootstrapping reproducible is a different problem from making everything reproducible. | 18:59:39 |
| Foxboron | Arch is the most reproducible still i think, but we have made no efforts into bootstrapping stuff | 19:00:22 |
| 3 Dec 2022 |
| cbwang | In reply to @foxboron:archlinux.org
Making the seed bootstrapping reproducible is a different problem from making everything reproducible. Yeah, so perhaps we can combine the package reproducibility of NixOS with the seed bootstrapping reproducibility of Guix | 00:28:28 |
| 6 Dec 2022 |
|  @federicodschonborn:matrix.org joined the room. | 13:01:02 |
|  @connorbaker:matrix.org left the room. | 15:17:32 |
| 8 Dec 2022 |
|  terrestrialpedestrian joined the room. | 18:04:54 |
| 10 Dec 2022 |
|  @rimuru:gentoo.chat changed their profile picture. | 00:25:49 |
| 12 Dec 2022 |
|  @notgne2:wizbos.club changed their profile picture. | 18:22:20 |
| 14 Dec 2022 |
|  @nahfe:nahfe.xyz changed their display name from NAHFE to Herman. | 10:55:05 |
|  Alesya Huzik joined the room. | 11:11:36 |
| @nahfe:nahfe.xyz changed their display name from Herman to NAHFE. | 11:31:35 |
| @nahfe:nahfe.xyz changed their display name from NAHFE to Herman. | 11:33:32 |
| 16 Dec 2022 |
|  fbecart joined the room. | 08:17:43 |
| 19 Dec 2022 |
| @rimuru:gentoo.chat changed their profile picture. | 11:54:08 |
| 20 Dec 2022 |
 tpw_rules | https://github.com/NixOS/nixpkgs/pull/206955 | 01:11:33 |
 @rick:matrix.ciphernetics.nl | Step by step | 09:43:54 |
| 24 Dec 2022 |
|  Jared Mohammed changed their profile picture. | 20:53:49 |
| 25 Dec 2022 |
|  @ahsmha:matrix.org left the room. | 10:37:15 |
| 26 Dec 2022 |
|  lassulus joined the room. | 11:25:19 |
 raboof | updated https://arnout.engelen.eu/nixos-iso-minimal-r13y/ - python looking healthier, one new issue: https://github.com/NixOS/nixpkgs/issues/207841 | 16:17:11 |
|  @jcmdln:matrix.org joined the room. | 18:27:16 |
| 27 Dec 2022 |
|  auk joined the room. | 21:31:54 |
| 28 Dec 2022 |
| @jcmdln:matrix.org | Are there automated tests for reproducing all Nix packages, or would I be correct in assuming that the focus is specifically for the packages included in the minimal iso? Out of curiosity I tried reproducing a few packages like postfix which is non-deterministic but I'm not sure if it should be reported. | 16:28:19 |
 @trofi:matrix.org | I'd say all the packages should be reproducible and it's worth reporting any non-determinism. | 16:30:11 |
