| 8 Jul 2024 |
|  @zzantares:matrix.org changed their display name from zzantares to demo-reset. | 06:00:25 |
| @zzantares:matrix.org removed their profile picture. | 06:00:33 |
|  @adham-omran:matrix.org left the room. | 06:33:29 |
|  Pyrox [Fox/It/She] changed their display name from Pyrox [Fox/It/She/They] to Pyrox [Fox/It/She]. | 06:53:22 |
|  @elonsroadster:matrix.org left the room. | 19:53:29 |
| 9 Jul 2024 |
 Null_A | does pgtk work on emacs? | 07:01:18 |
| Null_A | * does pgtk work on macos | 07:01:36 |
|  offset---cyan joined the room. | 10:50:43 |
 JoelMcCracken | idk a bit of ddging found this: https://www.reddit.com/r/emacs/comments/q1mxpz/pgtkgcc_emacs_on_macos_using_nix/ | 16:03:09 |
| JoelMcCracken | but I kinda mostly doubt it, or well, if it does, the emacs window would be an x/gtk based window not a macos native window | 16:04:44 |
| JoelMcCracken | i was wondering, i've been considering tyring to integrate org mode and nix so that my nix setups and evaluations are declarative. Does anyone know of any prior art/information on this? | 16:09:42 |
| JoelMcCracken | a concrete example of what i'm thinking is having an org mode file that tangles all the nix files in a derivation and has some other blocks that runs tests etc against it | 16:11:19 |
| JoelMcCracken | the main reason is i'm still working on learning nix and thinking that something like might help me be more explicit/understand what I is going on better. | 16:12:03 |
 bandithedoge | In reply to @jasonjckn:matrix.org
does pgtk work on emacs? why do you need pgtk on mac? | 16:22:25 |
| bandithedoge | In reply to @JoelMcCracken:matrix.org
i was wondering, i've been considering tyring to integrate org mode and nix so that my nix setups and evaluations are declarative. Does anyone know of any prior art/information on this? my idea would be to write a script that tangles everything to a temporary directory with a flake.nix and then points nixos-rebuild (or whatever you're using) to it | 16:24:22 |
| JoelMcCracken | yea that's kinda what i'm thinking | 16:24:47 |
| bandithedoge | you can use a nix-shell shebang line to get a minimal emacs as a dependency | 16:25:37 |
| JoelMcCracken | the primary issue with doing that that I see (for me) is that it would be error prone to set up (afaik, maybe there's a way to do it); ideally I would want to hack on the org babel tangling source stuff | 16:25:59 |
| bandithedoge | iirc tangling an org file is pretty quick | 16:27:14 |
| JoelMcCracken | yea it is | 16:27:55 |
| bandithedoge | you can make the directory non-temporary (~/.cache maybe?) and have the script check if it needs to be tangled | 16:27:56 |
| bandithedoge | * you can make the directory non-temporary (~/.cache maybe?) and have the script check if it needs to be re-tangled | 16:28:23 |
| bandithedoge | then you can just tangle from emacs and run the script when you need to rebuild | 16:28:58 |
| JoelMcCracken | so an upgrade to this I imagine would be like:
- a headline defines/groups a set of files into a specific nix directory+files content/experiment
- sub-headlines (or maybe just src blocks?) contain the source of the files for this headline's experiment
- data is saved to a dir named from the hash of the file names+contents so that it can be found again/deduped (though maybe this is already kinda done by nix itself, I'm still not super duper knowledgeable about nix) ((actually... I do think this is maybe unnecessary? though with nix flakes, I would want to have some kind of a place where the lock file is stored (not in the org file...) but able to be inspected/etc))
- once deriv is built, ideally can define some code to run against/explore that build (simple version would just be running shell in a nix shell for that deriv to inspect its output etc)
| 16:35:17 |
| JoelMcCracken | bc the main issue I forsee with this is the difficulty of setting up a new experiment based upon an old one... what i'm suggesting here would be "competing" with (the regular "have a project, edit the files and poke at stuff until it does what you want") | 16:36:49 |
| JoelMcCracken | idk maybe flake lock could be cached instead | 16:38:00 |
| JoelMcCracken | other thing about each "iteration of an experiment" having its own dir tho that I like is that I could cd into it and poke at it if I need to | 16:38:39 |
| JoelMcCracken | anyway thanks at the very least, glad the idea makes sense to someone else | 16:40:27 |
| JoelMcCracken | * anyway thanks at the very least for talking it through with me, glad the idea makes sense to someone else | 16:40:44 |
 bestlem | JoelMcCracken: There is https://github.com/rasendubi/dotfiles for everything generated for .org files | 19:42:53 |
