| 11 Dec 2024 |
 lillecarl | White demands too much of the wearer, I like being a slob 🤪 | 10:32:34 |
 markuskowa | White would be too much 😂 | 17:34:34 |
| 12 Dec 2024 |
| markuskowa | Just as a reminder: we have a little get-together next week Wednesday at 18:00. | 09:45:53 |
| 13 Dec 2024 |
| lillecarl | Found a flake related issue on the cheap-o Chromebook I bought. eMMC is being trashed pretty hard | 18:34:25 |
| lillecarl | Which is a flakes issue really, since they evaluate from store rather than FS | 18:34:59 |
| lillecarl | * Which is a flakes issue really, since they evaluate from store | 18:37:34 |
| 15 Dec 2024 |
| markuskowa | In what sense does the evaluation affect the store under flakes? | 19:08:22 |
| lillecarl | With flakes nix copies your git repo into the store and evaluates from there | 19:09:00 |
| lillecarl | So you're trashing your disk pretty hard, even though my repo is small. If you're on MMC it seems to break some regions of the block device | 19:09:42 |
| lillecarl | I'm gonna power wash the machine and see if that helps | 19:09:53 |
| lillecarl | But it's the same as running NixOS on a RPI with a shitty SD card, it'll die sooner than later | 19:10:20 |
| markuskowa | You mean, because you download a nixpkga tarball too often? | 19:10:27 |
| lillecarl | No the local git repo | 19:10:38 |
| lillecarl | ~/Code/nixos | 19:10:42 |
| markuskowa | Ah, ok I see. | 19:10:53 |
| lillecarl | But it might be the filesystem metadata blocks that get broken, I don't really know | 19:11:24 |
| lillecarl | They'll be written quite a lot with nixos/home-manager in general 😄 | 19:11:56 |
| markuskowa | But this is always the case, when you use "src = ./.", even without flakes. | 19:12:20 |
| lillecarl | Yeap, and they're still not implemented like gnu coreutils cp | 19:13:01 |
| lillecarl | If they were reflink copied on filesystems that support it, it would be so much lighter on block devices | 19:13:30 |
| lillecarl | (XFS, btrfs, bcachefs and friends would then just make a reflink and CoW if it's ever written to, but since it's a RO fs it's never written to so CoW only happens when you change a file within your repo) | 19:14:21 |
| lillecarl | But it's not implemented because it isn't portable to MacOS and every FS etc.... 😥 | 19:14:42 |
| markuskowa | Yes, reliying on fs specific features sounds tricky. | 19:17:41 |
| lillecarl | https://github.com/coreutils/coreutils/commit/25725f9d41735d176d73a757430739fb71c7d043 I mean, gnu coreutils does it by default for cp | 19:18:39 |
| markuskowa | Did you check if Lix has attempted to tackle this problem? | 21:00:18 |
| lillecarl | I did read the changelog for the 2 Lix releases that's been released so far but there was nothing that I could find. I kinda neglected the problem and spun up a hetzner VM to work through instead | 21:03:33 |
| 16 Dec 2024 |
 Claes | Unfortunately I can't make it this time. Got a cold and many end-of-year things piling up. Need to prioritize those. Hope you have a good time and see you next year instead! | 11:33:49 |
 igorramazanov | Ok, sorry for the late notice, but I secured an office space for us, so we don't have to hang out in to lobby, if we want to | 12:39:13 |
| igorramazanov | The office is in the "A-House" coworking building, near to Rådsmansgatan and Tekniska Högskolan stations, Östermalmsgatan 26A
If we want, then we can gather there | 12:41:39 |
| markuskowa | Sorry to hear. Get well and enjoy the holidays. | 12:50:26 |
