| 11 Jun 2024 |
K900 | I'm actually potentially interested in something like this | 07:43:42 |
K900 | But extremely far down the yak stack | 07:43:47 |
jade_ | In reply to @irenes:matrix.org which is a fair position but like... the world does have state. I don't think it's really right to build an abstraction that says there's never state. I think there's a missing counterpart to nix that deals with exactly the things nix doesn't - stuff that shouldn't be version-controlled or multi-homed for security reasons; backups; figuring out what you even have... oh, hi raitobezarius | 07:43:56 |
@irenes:matrix.org | oh, yes. this is many yaks in. | 07:44:11 |
@irenes:matrix.org | lol | 07:44:14 |
@irenes:matrix.org | there are a ton of things in lix which are way more important than this, anyway | 07:45:12 |
@irenes:matrix.org | but flakes broke my previous solution and working around it has been an operational nuisance for most of a year now so I was finally getting around to digging into it heh | 07:45:37 |
jade_ | you could just, not use flakes tbh. the nix3 cli is reasonably ok at not-flakes, the only really bad thing is nixos-rebuild, but that thing boils down to approximately two commands | 07:50:49 |
jade_ | * you could just, not use flakes tbh. the nix3 cli is reasonably ok at not-flakes, the only really bad thing is nixos-rebuild, but that thing boils down to approximately two commands and is thus pretty easy to replace | 07:50:58 |
@irenes:matrix.org | no, yeah, I mean | 07:51:20 |
@irenes:matrix.org | it's a real option | 07:51:22 |
@irenes:matrix.org | I just feel like that's punting on the problem | 07:51:32 |
@irenes:matrix.org | if I believe in flakes as, like, a good idea that should be saved | 07:51:41 |
@irenes:matrix.org | then I want this sort of thing to be possible in them | 07:51:48 |
@irenes:matrix.org | even if it has to be something special in the machine config to allow the specific exception or something, that's fine for my needs | 07:52:10 |
@irenes:matrix.org | in fact that's probably ideal from a security perspective | 07:52:17 |
@irenes:matrix.org | if I don't believe in flakes then I'm against flakes and I want to remove them lol | 07:52:38 |
@irenes:matrix.org | but either way I don't get to solve the problem just for myself | 07:52:50 |
@irenes:matrix.org | but on balance I do believe in flakes, just, that's my thought process | 07:56:25 |
jade_ | i don't think it should be saved personally. i think that most of it should be burned down and rewritten and relegated to a secondary component. we wouldn't drop support, we would just go put it in a corner where it doesn't need a lot of maintenance. the good features would be generalized, and the good ideas taken to make something new of the ashes.
flakes are a cute little lock file manager, and the schema would be a lot more useful if flake = true; inputs actually supported putting my own nixpkgs with overlays into them. if i could actually use the schema, it would be useful. but such as it is. a lot of the fancy flakes libraries tend to make it all a lot worse by forcing all the inputs unnecessarily, causing gratuitous downloads and making it impossible to stub inputs (cough flake-parts).
| 07:57:16 |
@irenes:matrix.org | I respect taht position | 07:57:29 |
@irenes:matrix.org | sp | 07:57:31 |
@irenes:matrix.org | perhaps in three more years of disillusionment I'll be in the same place | 07:57:39 |
jade_ | my "not using flakes" sure involves a lot of using flakes, i will say :P | 07:58:01 |
@irenes:matrix.org | I noticed that =p | 07:58:10 |
jade_ | i just write flakes that don't actually use any flake features except the locking | 07:58:20 |
@irenes:matrix.org | fwiw I am all for, like, reforging it in flame lol | 07:58:22 |
@irenes:matrix.org | any necessary changes to turn it into a thing people are actually happy with | 07:58:34 |
jade_ | * i just write flakes that don't actually use any flake features except the locking and purity, because i actually like those | 07:58:54 |
@irenes:matrix.org | but I do use devShells and templates and machine configurations | 07:59:00 |