| 20 Mar 2026 |
 piegames | (Nix is real and can't hurt you) | 14:06:14 |
| piegames | * (Nix is not real and can't hurt you) | 14:06:26 |
 emily | yeah I think it probably involved a \ or something in there too. but no REPL to jog my memory right now. | 14:07:27 |
| piegames | admittedly some of these are a bit tricky even when a better alternative exists, so we might at least improve the error message to point users in the right direction | 14:11:19 |
| piegames | I'd so much love to scrap all of this bullshit and redo a proper string syntax, but alas this will have to wait until langver which is soft-blocked on bytecode and Rust | 14:11:53 |
 KFears 🏳️⚧️ (they/them) | Why is it soft-blocked on bytecode and Rust? Because the current thing is a horrible pain? | 14:12:43 |
| piegames | the current deprecations are a best effort attempt at removing a loaded gun from users' hands | 14:12:54 |
| piegames | basically, yes. I tried 1.5y ago at https://gerrit.lix.systems/c/lix/+/2359/1 and eventually gave up. Having to create new, versioned, AST nodes with different semantics etc was inacceptably bad, and would have made maintenance in the long term even worse | 14:14:25 |
| piegames | so better to do it right from the beginning, clean up the codebase, put the correct abstraction layers in place, and then everything is much less unpleasant | 14:14:59 |
| KFears 🏳️⚧️ (they/them) | Fair enough | 14:17:02 |
| KFears 🏳️⚧️ (they/them) | We wouldn't attempt the C++ maneuver with a gun pointed to our head, so it's extremely impressive you tried that | 14:17:44 |
 whispers [& it/fae] | langver being the ability to experiment with changing the nix language while still being compatible with nix proper? (akin to rust's editions?) | 14:19:32 |
| whispers [& it/fae] | * | 14:20:09 |
 K900 | Basically that, yes | 14:20:19 |
| piegames | In reply to @kfears:matrix.org
We wouldn't attempt the C++ maneuver with a gun pointed to our head, so it's extremely impressive you tried that "How bad could it be?" | 14:32:06 |
| piegames | OTOH I wouldn't be able to know where to go if I hadn't tried and failed that way | 14:32:35 |
|  Patrick changed their display name from patrickdag to Patrick. | 14:52:01 |
| emily | (FWIW since ''\' is apparently still accepted I'm not sure I know of any issues with the escapes) | 15:41:42 |
| emily | (though I wouldn't be super surprised if there's something with ''\$ or ''\\) | 15:42:08 |
 aloisw | The only ones that I know of you can't write any more are when every line starts with at least one space. | 17:17:47 |
| emily | hmm, that's unfortunate. allow ''\<space> too? | 18:01:35 |
| aloisw | This was specifically rejected in favour of warning the user that it's just not supported any more. | 18:03:38 |
| emily | I'm not sure a string literal syntax that can't be used to form all of the valid strings in the language is a move in a less busted direction… | 18:08:36 |
| emily | (or what the problem with ''\<space> is, if ''\' is okay) | 18:09:16 |
| piegames | In reply to @emilazy:matrix.org
(or what the problem with ''\<space> is, if ''\' is okay) The latter was unfortunately necessary for escaping a ' at the end of a ''-string, because ''' is already defined to mean '' | 18:43:17 |
| emily | I think it's necessary before the end in some cases too. (but ''\<space> is also necessary as well, clearly) | 18:44:13 |
| piegames | It's not strictly necessary, because ${""} exists and is, arguably, no more hacky | 20:03:45 |
| aloisw | With that you could do single quote at the end, too. | 20:55:10 |
| piegames | heh | 21:12:14 |
| piegames | I totally did not think of that | 21:12:20 |
