| 28 Jan 2026 |
 aloisw | In reply to @sofiedotcafe:matrix.org
what if they don't expose such At least home-manager and impermanence have entry points that work without flakes. No entirely sure if they're official, but I haven't seen them break either. | 09:43:11 |
 toonn | I do use flake-compat for some Flake-only project. | 09:59:46 |
 delroth | (this discussion is imo precisely why flakes has won and why any flakes alternative without explicit compatibility support for flakes is doomed to fail) | 10:03:44 |
| delroth | * (this discussion is imo precisely why flakes has won and why (again, imo) any flakes alternative without explicit compatibility support for flakes is doomed to fail) | 10:07:18 |
 ShalokShalom | In reply to @coca162:matrix.org
You can also take a look at https://nilla.dev/ as well for this Looks neat! Although I kinda agree, so much as I personally dislike Flakes, would backwards compatibility be crucial for its success, I think.
Thanks for showing us this! | 11:49:50 |
 piegames | In reply to @delroth:delroth.net
(this discussion is imo precisely why flakes has won and why (again, imo) any flakes alternative without explicit compatibility support for flakes is doomed to fail) I mean flake-compat exists, not sure why keep asking about compat tools that already exists | 11:52:33 |
| piegames | * I mean flake-compat exists, not sure why keep asking about compat tools that already exist | 11:52:41 |
| piegames | I first-class integration what people want? One could add a flake input type to npins probably if that's what people want | 11:53:31 |
| piegames | * Is first-class integration what people want? One could add a flake input type to npins probably if that's what people want | 11:53:46 |
| delroth | I mean, I know I would want it if I were to use npins, but I don't use npins other than when I contribute to a project that uses it | 11:54:11 |
| delroth | so I'm probably not the right person to ask, or at least not via point-questions like this since most likely this isn't the only feature that stops me from using npins personally | 11:54:53 |
| piegames | I mean npins is fairly feature complete by now and in need of a successor, so I'm interested in hearing your show-stoppers for using it | 11:56:13 |
 Sofie 🏳️⚧️ (she/her) | vs npins? | 11:56:33 |
| piegames | Fixes won't land in npins, but will help inform any efforts at building a successor | 11:56:58 |
 Coca | you can use with it npins just like sprinkles, it only handles the schema part of flakes, just like in sprinkles | 11:57:16 |
 Lotte (it/its)/Cinny (she/her) θΔ& | you’d use both together seemingly | 11:57:21 |
| Coca | * you can use it with npins just like sprinkles, it only handles the schema part of flakes, just like in sprinkles | 11:57:27 |
| Coca | * you can use it with npins, it only handles the schema part of flakes, just like in sprinkles | 11:57:43 |
| piegames | (I consider nilla et al merely alternatives to npins – same as npins to niv – and not successors, because they don't fundamentally improve upon the core formula) | 11:58:03 |
| Sofie 🏳️⚧️ (she/her) | so, Nilla vs Sprinkles | 12:05:20 |
| Sofie 🏳️⚧️ (she/her) | Why not combine forces? | 12:05:29 |
| ShalokShalom | In reply to @piegames:flausch.social
I mean npins is fairly feature complete by now and in need of a successor, so I'm interested in hearing your show-stoppers for using it I read they are no full replacement for flakes | 12:15:35 |
| Coca | pinning inputs and schemas are two pretty different things for a single tool to handle, so most focus on one or the other | 12:20:30 |
 niko ⚡️ | What do you think npins is lacking in? | 13:12:25 |
| niko ⚡️ | Or, well, what in your eyes warrants a successor. Because I can only think of inputs list vs lockfile not being separate files being a tiny bit annoying, but that's about it? Probably my usecases aren't sophisticated enough though | 13:13:44 |
| piegames | That's the main point, yes | 13:14:36 |
| piegames | But also, anything wanting ti replace flakes should also do transitive dependency resolution in some way | 13:15:00 |
| piegames | In reply to @coca162:matrix.org
pinning inputs and schemas are two pretty different things for a single tool to handle, so most focus on one or the other Yes but in the end we need a tool that combines both, one way or another | 13:15:23 |
| niko ⚡️ | But then we're stepping out from just input locking to full on dependency management, probably also requiring standardized way of handing inputs and stuff like that. On one hand I agree but on the other my immediate reaction is "that's out of scope and I like how every projects brings their own locks and I don't want to extend the scope" | 13:19:59 |
| Coca | It's almost certainly out of scope for npins, but hopefully at least for people who want to take a crack at making the combined thing they can use something like libnpins to at least not have to reinvent the wheel again | 13:24:05 |
