| 11 May 2026 |
 askosia 🏳️⚧️ | thank you so much! i'll take a look at the latter option, opening PRs tends to be my go to for just, helping others not run into the same issues down the line :p | 14:12:06 |
| askosia 🏳️⚧️ | would there be any unexpected consequences from just swapping this out:
`sha256 = builtins.convertHash {
inherit hash;
toHashFormat = "sri";
hashAlgo = "sha256";
};
`
for just sha256 = hash? from waht I can tell fetchurl already has native support for sha256:hex, so the conversion seems unnecessary
| 14:42:26 |
| askosia 🏳️⚧️ | mm, is there any way to do discord-like code block highlighting on matrix? | 14:42:55 |
 zoë (@blokyk) |
| 14:43:04 |
| askosia 🏳️⚧️ | * that aside, is there any way to do discord-like code block highlighting on matrix? | 14:43:08 |
| zoë (@blokyk) | * ```lang
```
| 14:43:17 |
| askosia 🏳️⚧️ | huh | 14:43:16 |
| askosia 🏳️⚧️ | ahh thank you | 14:43:21 |
| zoë (@blokyk) | * ```lang
code
```
| 14:43:24 |
| zoë (@blokyk) | nested markdown is hard ok x) | 14:43:34 |
| askosia 🏳️⚧️ | would there be any unexpected consequences from just swapping this out:
sha256 = builtins.convertHash {
inherit hash;
toHashFormat = "sri";
hashAlgo = "sha256";
};
for just sha256 = hash? from waht I can tell fetchurl already has native support for sha256:hex, so the conversion seems unnecessary | 14:43:32 |
| askosia 🏳️⚧️ | you're all good lmao | 14:43:51 |
| askosia 🏳️⚧️ | i remember having to deal with it when trying to make a quine | 14:43:59 |
| askosia 🏳️⚧️ | it was a time for sure | 14:44:05 |
| askosia 🏳️⚧️ | ...i wonder how hard it'd be to make a quine in Nix | 14:47:12 |
| 12 May 2026 |
 Maxisthespy | In reply to @alexfmpe:matrix.org
Yeah this. A lot of the types would end up being Map String Any I’m not sure I actually find this a compelling argument, all types in any system reduce down to primitives, the point of modern higher type systems is to let us draw distinctions between the same primitives based on function | 16:36:18 |
| Maxisthespy | I’m not arguing for someone to implement it I just don’t think the fact that most nix types wind up being some combination of lists maps and strings that much of a barrier to the usefulness of a type system | 16:37:05 |
 alexfmpe | A type system would be useful, it's just that it'd take a lot of upfront work for it to be useful in the non trivial bindings because half the place is playing fast and loose with types | 16:38:45 |
| alexfmpe | Yeah I can be a boolean, or a string, or a list of strings, or a record, or a function that returns a record | 16:39:16 |
| alexfmpe | How do you type that without first making it sane? | 16:39:28 |
| Maxisthespy | I would go for a top down approach based on interfaces, for example does this follow a module interface ie does it provide outputs with the following names and vaguely the right types, that atleast would give some better sense of typing I feel | 16:41:37 |
| Maxisthespy | Which I suppose might be a slightly different thing but it atleast sidesteps the problem a bit | 16:42:05 |
| Maxisthespy | https://github.com/ryanrasti/typenix this project has a rather interesting claim | 16:43:53 |
| 13 May 2026 |
|  brume changed their display name from brume to brum. | 09:44:34 |
| brume changed their display name from brum to bru. | 09:47:29 |
| brume changed their display name from bru to brume. | 09:47:48 |
|  @gretj:catgirl.cloud left the room. | 10:21:22 |
|  phebee joined the room. | 10:38:21 |
|  Laura joined the room. | 14:40:04 |
| 14 May 2026 |
|  @vision69:matrix.org left the room. | 03:29:30 |
