| 7 Sep 2023 |
 ckie (they/them; limited keyboard usage, voice preferred) | doom doesn't encourage this and nix does but there's a big Mess to resolve when we make them touch | 13:45:49 |
 JoelMcCracken | yeah | 13:45:57 |
| ckie (they/them; limited keyboard usage, voice preferred) | In reply to @JoelMcCracken:matrix.org
for what i had been doing before as I've been trying to make all these packages work, it involved basically re-running home manager over and over yeah for me this takes a while because my evals are quite big | 13:48:48 |
| ckie (they/them; limited keyboard usage, voice preferred) | so i usually eval at point in my org config | 13:49:10 |
| JoelMcCracken | one thing about working on nde directly now is that I can run the check thing to see how its working | 13:49:22 |
| ckie (they/them; limited keyboard usage, voice preferred) | a friend who wasn't on Nix yet at the time could just M-x package-install and that was like. wow | 13:51:06 |
| ckie (they/them; limited keyboard usage, voice preferred) | like i went "oh you should get $X" and they just ran a few things in emacs and got the package | 13:51:23 |
| ckie (they/them; limited keyboard usage, voice preferred) | even with vanilla doom that'd be what, a doom sync and restart | 13:51:53 |
| JoelMcCracken | weljl you can i think run M-x doom sync | 13:53:13 |
| ckie (they/them; limited keyboard usage, voice preferred) | there's hints of a mainframe ideology in there though; what happens with a few emacsen on different machines | 13:53:20 |
| ckie (they/them; limited keyboard usage, voice preferred) | * there's hints of a mainframe ideology in there though; what happens with a few emacsen on different machines? | 13:53:23 |
| JoelMcCracken | sure | 13:54:02 |
| ckie (they/them; limited keyboard usage, voice preferred) | * there's hints of a mainframe ideology in there though; what happens with a few emacsen on different machines? | 13:54:29 |
| ckie (they/them; limited keyboard usage, voice preferred) | if we had a imperative2nix that'd be really neat | 13:54:50 |
| JoelMcCracken | for sure | 13:54:58 |
| JoelMcCracken | like somehow a way to generate an overlay from the current running doom | 13:55:22 |
| ckie (they/them; limited keyboard usage, voice preferred) | what's actually missing for that? what if we put a Nix in the doom | 13:55:28 |
| ckie (they/them; limited keyboard usage, voice preferred) | yea | 13:55:29 |
| ckie (they/them; limited keyboard usage, voice preferred) | like nix-shell -p but doom calls it from inside its package manager | 13:55:44 |
| ckie (they/them; limited keyboard usage, voice preferred) | see: [RFC 0040] "Ret-cont" recursive Nix | 13:57:09 |
| ckie (they/them; limited keyboard usage, voice preferred) | vaguely. not sure if that's precisely the right one | 13:57:45 |
| ckie (they/them; limited keyboard usage, voice preferred) | yeah this isn't the right one | 14:00:01 |
| ckie (they/them; limited keyboard usage, voice preferred) | I think I was thinking of [RFC 0092] Computed derivations but it's been >a year, need to reread | 14:02:38 |
| JoelMcCracken | yeah i don't know | 14:03:33 |
| JoelMcCracken | i imagine it would work like:
- a user picks a package to install from a list
- nix installs it
- emacs loads it within the running process
| 14:05:37 |
| ckie (they/them; limited keyboard usage, voice preferred) | mhm, it's essentially tearing out the parts of straight&emacs that do network I/O into their own emacs, running inside a nix build | 14:06:42 |
| ckie (they/them; limited keyboard usage, voice preferred) | which works happily for the top-level imperative emacs usecase (editing your config before rebuilding) but you want to have all of these prefetched in the build process once you're ready to deploy it to your 'prod' machine | 14:07:22 |
| ckie (they/them; limited keyboard usage, voice preferred) | recursive nix isn't actually a hard dependency-- we can crudely and inefficiently emulate it, but it would make this much more streamlined since it would run through approximately the same codepath in both imperative/Nix sandbox | 14:08:44 |
| ckie (they/them; limited keyboard usage, voice preferred) | * recursive nix isn't actually a hard dependency-- we can crudely and inefficiently emulate it, but it would make this much more streamlined since it would run through approximately the same codepath in both the imperative/Nix sandbox usecases | 14:08:49 |
| JoelMcCracken | yeah | 14:08:55 |
