| 12 Sep 2022 |
 k0kada (he/him) | I started rewriting a few months ago but have not found time to finish it | 17:12:21 |
 ckie (they/them; limited keyboard usage, voice preferred) | i actually don't use flakes | 17:12:35 |
| k0kada (he/him) | In reply to @ckie:ckie.dev
i see a convert-everything-to-module
Yeah, this was this rewriting that didn't go well :D | 17:12:38 |
| ckie (they/them; limited keyboard usage, voice preferred) | too restrictive | 17:13:05 |
| ckie (they/them; limited keyboard usage, voice preferred) | almost everything is homegrown | 17:13:23 |
| ckie (they/them; limited keyboard usage, voice preferred) | theres just niv and the insecure secrets module i stole from xe and fixed | 17:13:36 |
| k0kada (he/him) | In reply to @ckie:ckie.dev
too restrictive Before flakes I used to simply pin everything for each commit, but this started to get really tiring | 17:14:08 |
| ckie (they/them; limited keyboard usage, voice preferred) | the dialect has gotten advanced enough that this whole service is just 32 lines | 17:14:16 |
| k0kada (he/him) | I did took a look at niv, but eventually just decided to go full flakes | 17:14:35 |
| ckie (they/them; limited keyboard usage, voice preferred) | i think flakes were a real shame | 17:14:47 |
| ckie (they/them; limited keyboard usage, voice preferred) | sticking all the New Good Things in one thing was so stupid | 17:15:05 |
| ckie (they/them; limited keyboard usage, voice preferred) | * sticking all the New Good Things in one thing was so stupid in hindsight | 17:15:12 |
| ckie (they/them; limited keyboard usage, voice preferred) | like i occasionally get impurities and i have to weed them out manually because nix doesn't have a pureScope thing | 17:15:51 |
| ckie (they/them; limited keyboard usage, voice preferred) | and it's nice to have a standard flake.lock but something like niv or whatever should've been softly adopted as The Official Solution perhaps | 17:17:02 |
| k0kada (he/him) | Well, I miss a few things about flakes, but still like the idea in general | 17:17:08 |
| ckie (they/them; limited keyboard usage, voice preferred) | and the standardized outputs are frankly a joke but the inputs part and the entrypoint spec is great and should've also been its own thing | 17:17:32 |
| k0kada (he/him) | One thing that I really miss is having an input that doesn't propagate | 17:17:37 |
| k0kada (he/him) | Like dev inputs | 17:17:41 |
| ckie (they/them; limited keyboard usage, voice preferred) | the git tracking and source copying is absolutely stupid | 17:18:11 |
| ckie (they/them; limited keyboard usage, voice preferred) | at least the impl, the idea has some merit | 17:18:27 |
| ckie (they/them; limited keyboard usage, voice preferred) | like default.nix: import ./foo.nix, foo.nix empty staged file, unstaged changes with actual nix code | 17:19:11 |
| ckie (they/them; limited keyboard usage, voice preferred) | and it still runs | 17:19:16 |
| k0kada (he/him) | In reply to @k0kada:matrix.org
Actually, maybe I should just add it to the module Well, another idea to do in NixCon Hackathon | 17:22:38 |
| k0kada (he/him) | That or finally create a proper RetroArch module | 17:22:47 |
| k0kada (he/him) | Or both :D | 17:22:55 |
| ckie (they/them; limited keyboard usage, voice preferred) | i'd like nix to declare everything | 17:23:25 |
| ckie (they/them; limited keyboard usage, voice preferred) | but rebuilds are so damn slow | 17:23:28 |
| k0kada (he/him) | Yeah, it would be really nice if they could optimize the interpreter | 17:24:28 |
| ckie (they/them; limited keyboard usage, voice preferred) | it's already had its insides squeezed | 17:24:50 |
| k0kada (he/him) | I remember that I saw some years ago someone trying to do this, but I don't think they got really far | 17:24:55 |
