21 Oct 2024 |
KFears (annoying) | Fundamental ones, too | 20:54:42 |
KFears (annoying) | Well, I see it more as an official death penalty to CppNix rather than something that actually changes things, so that's just something that was already in the air since Eelco wasn't permabanned from the community, just not made official | 20:56:51 |
KFears (annoying) | Also, bending over backwards to not break GPL by "supervising regular daemon" (whatever that means) sounds like a lot of fun | 20:58:00 |
KFears (annoying) | Especially since CppNix isn't exactly known for being well-integrated into, well, anything | 20:58:22 |
aidalgol | In reply to @jakehamilton:auxolotl.org Is there someone I can DM about the bagel shaped thing that isn't a bagel? (Wanted to share something, but not sure if y'all are wanting stuff public) Thank you for putting donuts into my head. Now I want donuts... | 21:22:03 |
aidalgol | Unfortunately, there are no good donuts in this country. 😩 | 21:22:34 |
uep | In reply to @kfears:matrix.org Also, bending over backwards to not break GPL by "supervising regular daemon" (whatever that means) sounds like a lot of fun I mean it's one way to slowly start moving functions from the original codebase to the proxy? | 22:09:15 |
KFears (annoying) | In reply to @uep:matrix.org I mean it's one way to slowly start moving functions from the original codebase to the proxy? I mean, yes, I think that's the point, but I'm actually not really sure how it would look considering CppNix's very... reluctant integration history, I suppose? It's quite hard to extend what it does without linking against it | 22:12:38 |
uep | yeah, it's one way; not necessarily the way you'd choose without other considerations weighing in (like licensing, or politics) | 22:22:14 |
uep | RIIR in the same executable is a different set of tradeoffs | 22:23:10 |
KFears (annoying) | You can't have a closed-source binary linking to CppNix due to GPL licensing, so maybe I'm just not well-versed in the exact thing you're talking about, but it seems extremely hard to do | 22:24:11 |
uep | no, that's exactly what I mean | 22:24:35 |
uep | if you want to migrate from an older to a newer codebase, there are a few ways to do things | 22:25:02 |
uep | licensing isssues like that are one factor pushing that choice in a particular direction | 22:25:28 |
uep | but even without that, writing façade that slowly absorbs the things it's proxying is a valid technical strategy | 22:26:10 |
KFears (annoying) | Oh yeah, definitely | 22:26:57 |
KFears (annoying) | Like yeah, I was trying to say something closer to "with how hard CppNix is to work with having access to the code and having interfaces for high-level actions, trying to do low-level actions without the code is gonna be hell" | 22:28:08 |
uep | yeah, some things are easier than others; being a proxy to all the caches so you can manage config and trust and discovery separately from cppnix's mechanisms is one of the easier ones, apparently. they're doing that, stuff like peerix tried too | 22:30:46 |
KFears (annoying) | I think that's pretty tame for functionality, tbh | 22:31:22 |
uep | yeah | 22:31:30 |
KFears (annoying) | I think they also have like authentication for flakes with RBAC and that just sounds weird | 22:32:08 |
uep | if your goal is to rip things out of the core and simplify it, vs just pave over the complexity | 22:32:55 |
KFears (annoying) | Maybe they like authentication in their separate daemon and download the thing after authenticated and then feed it into CppNix with its low-level commands? That's how I would approach the problem | 22:33:00 |
KFears (annoying) | * Maybe they like do authentication in their separate daemon and download the thing after authenticated and then feed it into CppNix with its low-level commands? That's how I would approach the problem | 22:33:08 |
KFears (annoying) | This makes sense with all the marketing around flakes in particular, too, but by that point I'm personally wondering why would I even be interested in Nix | 22:33:52 |
KFears (annoying) | RBAC and authentication would be nice, but I'd imagine them happening at the Nix-daemon level, not with flakes | 22:34:55 |
KFears (annoying) | I'm just kinda confused at the whole value proposition here and I feel like the product being sold is different from the product being advertised, but what do I know | 22:36:57 |
remi-gelinas | Anyone know off the top of their head if there's an easy way to disable TLS verification for Lix in the docker container? | 23:36:01 |
remi-gelinas | I'm trying to build some stuff in a different platform container, but MITM proxies are blocking it. The cert is in my host system store, but not in the Lix docker container ofc | 23:36:39 |
hexa | has anyone noticed remote builds getting stuck until you login via ssh to the remote builder? | 23:47:25 |