| 24 Sep 2021 |
 David Arnold (blaggacao) | Updating inputs might always break something. This is because we opted to swap dependencies (a.k.a. "overlays" / "follows"). This reduces the amount of storage required on disk, but might break some things. | 13:24:48 |
| David Arnold (blaggacao) | NUR can just be dropped, it is not wired anywhere. It is just a relict of that inital intent in the template. | 13:25:47 |
 @kraftnix:matrix.org | well the nur output is still technically used + the overlay added in devos | 13:26:36 |
| David Arnold (blaggacao) | "overlays" / "follows" are two approaches to the same goal: homologate inputs, and (eventually) "re-evaluate" the world in order to re-use dependencies. | 13:27:14 |
| David Arnold (blaggacao) | If not, ome can end up with 10 different bash binaries used in different parts and things like that, heavy on disk. | 13:27:47 |
| @kraftnix:matrix.org | if the intent is to drop it I would support removing it from devos + documenting it as an example of how to easily add extra inputs with their overlays and use packages from them | 13:28:05 |
| David Arnold (blaggacao) | * If not, one can end up with 10 different bash binaries used in different parts and things like that, heavy on disk. | 13:28:06 |
| David Arnold (blaggacao) | kraftnix I support that. We can re-evaluate the intent later on in case we manage to do a tighter integration with the NUR ecosystem. | 13:29:04 |
| David Arnold (blaggacao) | * kraftnix I support that. We can re-evaluate the intent later and in case we manage to do a tighter integration with the NUR ecosystem. | 13:29:22 |
| @kraftnix:matrix.org | In the meantime (until nur is dropped) should we add in the actual nur input to avoid unexpected breakages? or would you be more in favour of just removing it outright immediately? | 13:30:15 |
| David Arnold (blaggacao) | It's a template, not even a ref implementation. Let's jut drop it. | 13:30:51 |
| David Arnold (blaggacao) | Btw. I'd regard the examples folder in digga more a ref implementation of some of the digga features. | 13:32:11 |
| David Arnold (blaggacao) | I think we somehow need to communicate that better. | 13:32:22 |
| @kraftnix:matrix.org | ah, I had not spotted that addition to digga, that's neat | 13:32:53 |
 ultranix | i still dont understand still why i cant rename my emacs user profile directory emacs and i have to make it emacs-ng or something different, or i get the error
The option `home-manager.users.janus.inputs' does not exist. Definition values:
- In `/nix/store/6xm5sn1m8bpgqb3j6g2hljvz7lc3l65i-source/users/janus': { }
| 18:47:51 |
| ultranix | * i still dont understand still why i cant rename my emacs user profile directory emacs and i have to make it emacs-ng or something different, or i get the error
The option `home-manager.users.janus.inputs' does not exist. Definition values:
- In `/nix/store/6xm5sn1m8bpgqb3j6g2hljvz7lc3l65i-source/users/ultranix': { }
| 18:47:55 |
| ultranix | * i still dont understand why i cant rename my emacs user profile directory emacs and i have to make it emacs-ng or something different, or i get the error
The option `home-manager.users.janus.inputs' does not exist. Definition values:
- In `/nix/store/6xm5sn1m8bpgqb3j6g2hljvz7lc3l65i-source/users/ultranix': { }
| 18:48:17 |
| ultranix | the only thing that's different is the directory name, and update that name in the suites | 18:48:57 |
| ultranix | ive tried literally emptying the emacs/default.nix file as well | 18:54:09 |
| ultranix | i'm sure it has to do with mkFlake, maybe because the directory/profile name in home.importables.suites (emacs) is the name name as a flake input/output? | 18:59:28 |
| ultranix | i have emacs.url = "github:nix-community/emacs-overlay" as an input | 19:00:03 |
| ultranix | as a matter of fact that's what it is. i think digga has a bug, has anyone tried something similar like this? | 19:03:14 |
| ultranix | i just removed my emacs input/output and kept the emacs user profile in mkFlake and it works as expected | 19:04:33 |
| ultranix | * i'm sure it has to do with mkFlake, maybe because the directory/profile name in home.importables.suites (emacs) is the same name as a flake input/output? | 19:04:54 |
| ultranix | i mean except now i cant use the nix-community/emacs-overlay packages | 19:05:32 |
| ultranix | i could keep the names different i suppose | 19:07:14 |
 aciceri | ultranix: I just tried to add Emacs as you putting a profile inside the users/ folder and I confirm your problem (and workaround i.e. different names).
However I can't help. | 19:35:18 |
| aciceri | But I take the opportunity to ask how to "disable" a mkFlake's output. My use case is this: I added a NixOS configuration for my aarch64 notebook and when I use nix flake check on my pc (which has crosscompiling enabled) it correctly runs but when I put this command in a Github workflow it tells me that "a 'aarch64-linux' with features {} is required to build..." (since the Github runner is an x86_64 machine).
So how can I tell mkFlake/digga to do not generate checks for a particular platform? Any workaround? | 19:40:14 |
| David Arnold (blaggacao) | That interface sits really at the nix level I guess. You could just override the checks output. | 19:41:32 |
| David Arnold (blaggacao) | I don't understand the gist of the "emacs"-issue, though. Can you boil it doen? | 19:42:17 |
