| 30 Aug 2023 |
 Mic92 | Also makes you suspicious in field operations, I think. I think well up-to-date phones are also not easy to hack by normal average organisations. They all have signed bootloader and apps run in sandboxes. Maybe turn off stuff like bluetooth as well and have a sim card that only does internet. | 07:27:36 |
| Mic92 | * Also makes you suspicious in field operations, I think. I think an up-to-date phones is also not easy to hack by normal average organisations. They all have signed bootloader and apps run in sandboxes. Maybe turn off stuff like bluetooth as well and have a sim card that only does internet. | 07:28:12 |
| 31 Aug 2023 |
|  moritz.hedtke removed their display name moritz.hedtke. | 16:13:23 |
| 1 Sep 2023 |
 woobilicious | Yeah, phones are surprisingly well locked down, too bad that Android is trash for power users. | 04:32:30 |
| woobilicious | Heads is really good against evil maid, better than bitkeeper. | 04:34:13 |
| 3 Sep 2023 |
|  Khakish joined the room. | 19:40:48 |
| 7 Sep 2023 |
| woobilicious | Mic92: were you maintaining the bcachefs stuff in nixpkgs? I've kinda distanced myself from upstream project, and the code is very stable now, so I was wondering if I should put some effort in to just migrating my repo to nixpkgs. | 09:19:31 |
| Mic92 | woobilicious: I do. | 09:20:33 |
| Mic92 | I update bcachefs-tools and the kernel every now and than. | 09:21:14 |
| woobilicious | Yeah I borrowed some code...I'm not sure what nixpkgs rules are around external nix stuff. my goal was originally to keep the code inside bcachefs-tools. But Kent never takes the time to test it and it stops working occasionally, He's pretty open to take patches tho. | 09:23:36 |
| Mic92 | You mean licensing? | 09:42:27 |
| Mic92 | I think patches can be multiple licenses. Nixcode is MIT. | 09:42:45 |
| Mic92 | However if you wrote it, you are free to relicense under MIT. | 09:42:56 |
| Mic92 | * I think patches can be multiple licenses. Nix code is MIT. | 09:43:05 |
| woobilicious | hmm I wasn't thinking about licencing (though it's probably worth thinking about), more most nixpkgs packages use a nix file within nixpkgs, I never see it use an external .nix file from an upstream. | 09:46:06 |
| Mic92 | Ah. Yeah, we don't do that. | 09:46:50 |
| Mic92 | nixpkgs needs to be self-contained | 09:46:58 |
| Mic92 | Nix code needs to be copied, we don't import stuff | 09:47:13 |
| woobilicious | IMHO, I think your code is working fine for bcachefs-tools, only issue is I think I spotted a lack of hardware accelerated crc32c and I think it was because the nixpkgs code wasn't putting the correct modules in to the initrd. | 09:49:56 |
| woobilicious | couldn't rebase bcachefs on to 6.4.x and I've been running 6.4.0 for ages and I'm starting to suspect my kernel is bugging out, firefox kept crashing, and and updating it only replaced the crash with a infinite loop caused by a page fault in libc memmove, thank god 6.5 is out. | 15:49:20 |
| 9 Sep 2023 |
|  Moritz Sanft joined the room. | 12:14:38 |
| 10 Sep 2023 |
| Mic92 | Ok. I guess I have update nixpkgs than | 16:37:32 |
| Mic92 | Using bcachefs on a desktop is vrave | 16:38:10 |
| Mic92 | *brave | 16:38:17 |
| woobilicious | I've been using it since 5.9, haven't had much data loss yet...lol Kent has a good splattering of asserts so it usually just goes read-only if there's a bug, more likely to trigger asserts during fsck when they're meant to be off (so they can actually fix the error), than have data corruption. | 16:55:40 |
|  MoKe12g changed their profile picture. | 23:20:19 |
| 11 Sep 2023 |
| Mic92 | woobilicious: is this maybe the crc module missing? https://github.com/NixOS/nixpkgs/pull/252348#issuecomment-1712058174 | 15:21:16 |
| Mic92 | mhm but crc32 is probably not used for crypto stuff. | 15:21:58 |
| woobilicious | no crc32 is turned off with encrypted data, already have MACs | 21:21:51 |
| woobilicious | Mic92: ^ | 21:22:33 |
