| 14 Nov 2025 |
 raitobezarius | i already know what caused it, don't worry | 21:24:34 |
| 15 Nov 2025 |
|  suya changed their profile picture. | 11:16:28 |
 Emma [it/its] | i wonder if its possible to build less of lix | 14:34:04 |
| Emma [it/its] | ie. ommitting nix-channel etc | 14:34:36 |
 aloisw | Such a possibility does not exist right now. I'm also afraid it wouldn't help much, as nix-channel is not that large and mostly works by calling other stuff that has to be there anyway. | 14:48:16 |
|  flax changed their display name from Jean to flax. | 15:31:38 |
 Katie | Hello. I have a question about prior art in Nix language RFCs/pre-RFCs of if anyone had anyone had ever proposed something along the lines of Regex Literals?
Considering that the Lix project is interested in evolving the language12, I figured this might be a good place to ask?
(Wasn't sure if this moreso belonged here or in Development)
| 18:28:31 |
| Katie | * Hello. I have a question about prior art in Nix language RFCs/pre-RFCs of if anyone had anyone had ever proposed something along the lines of Regex Literals?
Considering that the Lix project is interested in evolving the language[1][2], I figured this might be a good place to ask?
(Wasn't sure if this moreso belonged here or in Development)
| 18:29:12 |
 K900 | Was never proposed AFAIK and also Nix regex semantics are wildly underspecified | 18:31:29 |
| K900 | @aloisw has horror stories if you want | 18:31:34 |
| aloisw | The "horror stories" I found are mostly about finding a library that actually works. | 18:34:24 |
| aloisw | Yes, such a library exists (TRE with proper configuration), but never got around to rejecting its extensions. | 18:35:00 |
| aloisw | Regex literals are also something that looks more related to URL literals than to regex implementations, and maybe it suffices to say that the former were thrown out some time ago. | 18:35:47 |
| K900 | Actually now that I think about it | 18:36:14 |
| K900 | What would a regex literal even do | 18:36:19 |
| K900 | Like would it have function semantics? | 18:36:25 |
| K900 | So /foo/ bar desugars to builtins.match "foo" bar? | 18:36:42 |
| aloisw | no, it would just be another way of writing a string to confuse users | 18:36:49 |
| Katie | Oh that's an interesting thought | 18:36:50 |
| K900 | Cause I kinda hate that | 18:36:52 |
 522 it/its ⛯ΘΔ | i assume url literals were at least checked for validity? | 18:37:15 |
| K900 | Not really | 18:37:23 |
| K900 | And "validity" for URLs is a very nebulous concept anyway | 18:37:32 |
| aloisw | nix-repl> /foo/ bar
error: path has a trailing slash
at «string»:1:6:
1| /foo/ bar
| ^
surprisingly enough that syntax is free
| 18:37:36 |
| 522 it/its ⛯ΘΔ | /foo[/ should fail to evaluate, at least | 18:37:40 |
| K900 | Most things you would not expect to be valid URLs actually are | 18:37:41 |
| Katie | ...this is roughly what I imagined, but you do make a good point 😅 | 18:37:44 |
 aftix | you can write regex in '' to avoid double backslashing already? | 18:37:54 |
| K900 | Yeah I think the question here really is, what kind of semantics do you want | 18:38:07 |
| K900 | And also why | 18:38:09 |
