| 16 Mar 2026 |
 piegames | what you're doing here is abusing an implementation detail that frankly shouldn't have existed in the first place, the only reason it's not deprecated is because it is frankly useful in the absence of saner alternative | 22:34:55 |
| piegames | * what you're doing here is abusing an implementation detail that frankly shouldn't have existed in the first place, the only reason it's not deprecated is because it is frankly useful in the absence of saner alternatives | 22:34:57 |
 raitobezarius | i also don't think it makes anyone a service to learn about how this API abuse technique | 22:35:38 |
| raitobezarius | if you use it for your local systems, why not | 22:35:50 |
| raitobezarius | if you build an entire framework around it, :D | 22:35:55 |
| piegames | to be fair, for __findFile we are unlikely to brick it before we can provide some alternative | 22:36:36 |
| piegames | (not for lack of trying lol) | 22:36:42 |
 zoë (she/her) | while making frozenpins i did wonder why they'd even add this in the first place, i'm not sure what the value was for the nix impl | 22:36:46 |
| piegames | (I tried to deprecated it in the past and the snix people complained with somewhat valid use cases using this hack) | 22:37:06 |
| piegames | For most of these primops the reason is "it was the path of least resistance" | 22:38:23 |
| zoë (she/her) | makes sense for a thesis project like this tbf, in a way it's a shame it had these growing pains that didn't get resolved | 22:39:53 |
| zoë (she/her) | * while writing frozenpins i did wonder why they'd even add this in the first place, i'm not sure what the value was for the nix impl | 22:40:02 |
 vczf | To be fair, the angle bracket hack is not at all necessary for den, and it's not built around it | 22:40:48 |
| zoë (she/her) | looking through these docs i ended up on vic/with-inputs, which is... basically a better version of what i lost my mind on during the weekend... one of these days i'll actually write an original tool x) | 22:42:08 |
| zoë (she/her) | at least it was fun to do :D | 22:42:51 |
| vczf | I gave den another look because my home-grown system for per-hosts/users config was impossible to port to nix-darwin | 22:44:11 |
| vczf | So I feel it too lol | 22:44:17 |
| 17 Mar 2026 |
| piegames | zoë (she/her): I asked our dragon and while there are ways of getting scopedImport caching to work, the real trouble is that it would break the language one way or another no matter how you implement it | 10:47:52 |
| piegames | Because either way it becomes a primitive for observing equalities of values in the language which should semantically not be exposed | 10:48:02 |
| piegames | so the only real solution is to gather your use cases and design something new which solves these issues without the hacks | 10:48:31 |
| zoë (she/her) | In reply to @piegames:flausch.social
so the only real solution is to gather your use cases and design something new which solves these issues without the hacks
gather your use cases
well, most of my uses of it have been cursed things like the <bracket> syntax thing, which is pretty unserious on my side. but i am curious: how legitimate/"serious" of a use case for you feel that is?
| 11:22:12 |
| piegames | Think higher level. Ask yourself "why" you need this three times and that's what we want to implement | 11:26:06 |
| zoë (she/her) | *
gather your use cases
well, most of my uses of it have been cursed things like the <bracket> syntax thing, which is pretty unserious on my side. but i am curious: how legitimate/"serious" of a use case for you feel that is? | 11:26:09 |
| piegames | Though most likely the answer will resolve around needing something like a proper dependency importing system | 11:26:39 |
 sterni | raitobezarius: seems like build03 is back in buildkite? | 12:11:42 |
 delroth | I'll go kill it before it causes too many problems | 12:19:41 |
| delroth | raitobezarius: can you deploy a proper disable until you have time to fix that machine? I don't think it's in afnix infra? | 12:20:09 |
| delroth | (oh, it is, I'll deploy that then) | 12:21:53 |
| raitobezarius | In reply to @delroth:delroth.net
(oh, it is, I'll deploy that then) Thanks | 12:23:03 |
|  @draupnir:lix.systems left the room. | 14:32:38 |
