| 18 Sep 2024 |
toonn | If you're only building and distributing you're not an end-user. | 13:05:01 |
emily | otherwise it's just an unambiguous copyright violation | 13:05:11 |
emily | FWIW it's fine in Nixpkgs because it's marked as unfree (not unfreeRedistributable) | 13:05:19 |
emily | but I guess something on the nix-community infra is putting it in a cache when it really shouldn' | 13:05:35 |
toonn | The copyright is on the source code though, not the binaries. | 13:05:36 |
emily | * but I guess something on the nix-community infra is putting it in a cache when it really shouldn't | 13:05:37 |
emily | yes, and binaries are derivative works of source code | 13:05:46 |
emily | and Nixpkgs patches the source before compiling, so it also prepares a derivative work of the source code in the process | 13:05:59 |
Kamilla 'ova | In reply to@emilazy:matrix.org but I guess something on the nix-community infra is putting it in a cache when it really shouldn't https://hydra.nix-community.org/build/899731 | 13:09:55 |
emily | right. maybe related to https://github.com/NixOS/nixpkgs/pull/342529 https://github.com/numtide/nixpkgs-unfree ? | 13:10:31 |
emily | building and redistributing unfreeRedistributable is already a minefield given the various complex use restrictions many of those licences have, but doing it for unfree is a terrible idea | 13:10:58 |
emily | In reply to @emilazy:matrix.org right. maybe related to https://github.com/NixOS/nixpkgs/pull/342529 https://github.com/numtide/nixpkgs-unfree ? indeed related https://hydra.nix-community.org/jobset/nixpkgs/unfree#tabs-configuration | 13:11:38 |
Kamilla 'ova | last time I checked nixpkgs-unfree (~1 year ago) aseprite was not in their cache | 13:11:44 |
emily | which is confusing given zimbatm's statement of "our intent here is not to redistribute" | 13:11:50 |
Kamilla 'ova | I think this is related to https://github.com/nix-community/infra/pull/1406 | 13:12:09 |
Kamilla 'ova | in my opinion, building all unfree packages is good, but publishing them in a public cache is bad | 13:13:22 |
emily | posted https://github.com/NixOS/nixpkgs/pull/342529#issuecomment-2358438965 | 13:13:48 |
emily | In reply to @kamillaova:matrix.org in my opinion, building all unfree packages is good, but publishing them in a public cache is bad for unfreeRedistributable it seems like a good idea if someone is willing to take on the risk. unfortunately lib.licenses.unfree often means "impossible to build without committing a copyright violation"; we have it overloaded with a hypothetical lib.licenses.illegal | 13:14:24 |
emily | so it's a quite significant risk to do it for lib.licenses.unfree even in the absence of redistribution | 13:14:36 |
Kamilla 'ova | I think adding redistributableOnly option to release-unfree.nix will be fine | 13:14:38 |
Kamilla 'ova | In reply to@emilazy:matrix.org posted https://github.com/NixOS/nixpkgs/pull/342529#issuecomment-2358438965 via Cachix
I think hydra.nix-community.org also provides a way to download the output but it throws an exception inside the hydra code (https://matrix.to/#/#nix-community:nixos.org/$I6ODgqJ-1MNHvgq_lBx2AtpUA-q2p1llHYHAyMzKlGg) | 13:18:46 |
Kamilla 'ova | via Cachix
I think hydra.nix-community.org also provides a way to download the output but it throws an exception inside the hydra code (https://matrix.to/#/#nix-community:nixos.org/$I6ODgqJ-1MNHvgq_lBx2AtpUA-q2p1llHYHAyMzKlGg) | 13:18:57 |
emily | right. well "via Cachix" is more concerning because it means that we're also inducing Cachix into copyright violations, which they presumably wouldn't be happy about :) | 13:19:29 |
Mic92 | Isn't cachix just a provider in this case? | 17:03:25 |
emily | they'd be obligated to act on it if informed at least :/ | 17:07:38 |
emily | I would not want to see the nix-community cache obliterated by a DMCA 🫠| 17:08:36 |
| moved to @amadaluzia:tchncs.de changed their profile picture. | 23:01:40 |
| 20 Sep 2024 |
zowoq | In reply to @zowoq:matrix.org Breaking changes for buildbot.nix-community.org: https://github.com/orgs/nix-community/discussions/1457 I'll do this buildbot update now. | 02:39:04 |
Austin Horstman | In reply to @zowoq:matrix.org I'll do this buildbot update now. Seeing a bunch of cached failures in buildbot today, like this https://buildbot.nix-community.org/#/builders/3228/builds/31, is there any easy way to link back to the original reported error? | 19:14:37 |
Austin Horstman | In reply to @zowoq:matrix.org I'll do this buildbot update now. * Seeing a bunch of cached failures in buildbot today, like this https://buildbot.nix-community.org/#/builders/3228/builds/31, is there any easy way to link back to the original reported error? Looking at the history for that https://buildbot.nix-community.org/#/builders/3122 I see it go from success -> canceled -> cached failure | 19:16:33 |