| 26 Apr 2025 |
 Fernando Rodrigues | Huh, you should probably ask for a nixos.org alias then. | 14:40:22 |
 Yorusaka Miyabi | In reply to @emma:rory.gay
in that case, the answer to your original question is no 👀 I think it's likely safe to do so, mainly 'cause that room is invite only, and only people having a GitHub account can be invited to this room (using a webservice currently protected with Anubis) | 14:45:55 |
 @emma:rory.gay | so how would i join without a github account | 14:46:27 |
|  debabrata pattnayak joined the room. | 16:11:45 |
|  haylin joined the room. | 17:56:33 |
| haylin changed their profile picture. | 18:03:20 |
 uep | Unfortunately, mjolnir does not have configuration options to allow selecting channels for the image protection, it's all it nothing | 22:47:28 |
| uep | * Unfortunately, mjolnir does not have configuration options to allow selecting channels for the image protection, it's all or nothing | 22:47:47 |
| uep | We want to upgrade to draupnir but I don't think that's any different in this regard. If it really becomes an issue for ongoing usage, we'd have to look at running multiple instances with different settings. | 22:50:16 |
 Cat | Emma [it/its]: how easy is it to run multiple instances of bot mode now again with the module you have been brewing? | 23:16:14 |
| @emma:rory.gay | just as easy as it would be with mjolnir | 23:16:34 |
| @emma:rory.gay | not a usecase i accounted for, its something thats been on my mind but didnt want to bother with | 23:17:06 |
| Cat | Cat is trying to say that as far as i understand under nix its apparently not too hard to host multiple copies of the bots. And there is always the Appservice mode. | 23:17:14 |
| @emma:rory.gay | AS mode isnt supported in my module ;) | 23:17:25 |
| @emma:rory.gay | but tbh in my own infra i just use nixos containers | 23:17:36 |
| @emma:rory.gay | https://cgit.rory.gay/Rory-Open-Architecture.git/tree/host/Rory-ovh/services/containers | 23:18:01 |
| @emma:rory.gay | possibly not the most efficient way to do it, but it sure does work well | 23:18:35 |
| @emma:rory.gay | though, the module isnt that complicated, so you could easilly just lib.map it to systemd services | 23:19:21 |
| @emma:rory.gay | especially if you dont bother with validation | 23:19:33 |
| uep | Yeah, I'd just use multiple containers for this | 23:26:12 |
| @emma:rory.gay | yeah, its what i do, the only annoying thing is that requires networking.nat | 23:26:41 |
| uep | It's the conduit maintenance for multiple instances that is the issue, not the deployment | 23:26:54 |
| @emma:rory.gay | at least if you do the me thing of adding all your nginx vhosts to your hosts file as 127.0.0.1 | 23:26:59 |
| @emma:rory.gay | tbh, dont do the me thing, abstract it into a function | 23:27:29 |
| uep | Oh I just give every container an IP, either on a bridge or a routed dinner | 23:27:33 |
| uep | * Oh I just give every container an IP, either on a bridge or a routed subnet | 23:27:51 |
| @emma:rory.gay | so do i, but it doesnt work without NAt enabled for some reason | 23:27:55 |
| @emma:rory.gay | didnt want to bother too much, my setup works (though i did break it once by trying to set up a wireguard server...), so for as long as it does i wont touch it | 23:28:29 |
| uep | * It's the config maintenance for multiple instances that is the issue, not the deployment | 23:28:43 |
| uep | In reply to @uep:matrix.org
It's the config maintenance for multiple instances that is the issue, not the deployment They can at least share ban lists via a channel, that's legally most of it | 23:29:30 |
