| 17 Apr 2026 |
 whispers [& it/fae] | i suspect a lot of it (e.g. this case of library loading) is mostly down to the fact that it's a lot of work to get right and so it was never implemented, but the effects are the same | 22:48:28 |
| whispers [& it/fae] | * | 22:48:56 |
| whispers [& it/fae] | * | 22:49:33 |
 figsoda | something like system-deps built into cargo would definitely be interesting to see | 22:58:21 |
 nim65s | do you think a bazel(-ish) layer would help ? | 23:06:06 |
 emily | unsandboxed proc macros were also a mistake :) | 23:25:05 |
| emily | but tbh cc crate is a worse offender than anything pkg-config-related | 23:25:17 |
| figsoda | agreed | 23:31:47 |
| 18 Apr 2026 |
 Tumble | are you supposed to put Cago.lock in gitignore? | 15:15:49 |
| Tumble | i was thinking of building this for myself
https://github.com/waywallen/waywallen/blob/main/.gitignore
but they gitignored the Cargo.lock file | 15:16:03 |
| Tumble | * are you supposed to put Cargo.lock in gitignore? | 15:16:16 |
 K900 | Generally no | 15:16:23 |
| K900 | Also holy shit is that overengineered | 15:17:33 |
| Tumble | why are most of the things i decide to compile myself, aways have a weird build setup | 15:18:15 |
| Tumble | will i have to generate the lock file seperatly and then slide it into the build proces | 15:20:46 |
| K900 | I mean I would recommend not running that thing at all | 15:21:44 |
| K900 | But that's probably not the answer you're looking for | 15:21:52 |
| K900 | But there are definitely better ways to do whatever the fuck that's doing | 15:22:02 |
| Tumble | i could do that if there were other projects that do the same thing perhaps | 15:23:57 |
| K900 | https://github.com/luisbocanegra/plasma-smart-video-wallpaper-reborn | 15:24:34 |
| 19 Apr 2026 |
 Ralith | I think something bazel-ish is absolutely necessarily the way forwards in the long run, leaning a bit more heavily into hermeticity and hopefully decrufted | 03:41:51 |
| Ralith | I hoped buck2 would be this but meta dgaf about external users | 03:42:13 |
| Ralith | fingers crossed for https://vixen.rs/ | 03:42:19 |
| Ralith | in the meantime if you need to actually get some work done, bazel is Fine if you take care. I think it's less idiosyncratic than, like, Nix, though a lot easier to shoot yourself in the foot with hermeticity-wise. | 03:46:49 |
| Ralith | I think this specifically is overstated. There's a lot of folks successfully using Rust without cargo. It's some extra work, but it's typically easier than packaging stuff for nixpkgs. The rust/cargo folks have done a good job of making this friendly to automation.
Notably, it is much easier to port 99.9% of Rust crates to a new build system than to do the same for your average C/C++ project.
| 03:53:35 |
 aphasic | In reply to @tumble1999:matrix.org
i could do that if there were other projects that do the same thing perhaps considering it leans on other programs for everything except kde compatibility im near certain there are | 04:43:03 |
 Benedikt | Is it realistic that nixos will switch too a package per crate model like guix did? | 10:06:16 |
| K900 | It's been proposed before but it requires a huge amount of work | 10:16:00 |
| 20 Apr 2026 |
|  Four changed their display name from bivsk to Four. | 14:16:31 |
| 23 Apr 2026 |
 ghpzin | There might be a new problem with fetchCargoVendor - crates.io started blocking current python requests user-agent python-requests/2.33.1 that is used in fetch-cargo-vendor-util.py (because it does not set one explicitly in script):
curl -v https://crates.io/api/v1/crates/aho-corasick/1.1.4/download -L -A "python-requests/2.33.1"
< HTTP/2 403
...
{"errors":[{"detail":"We are unable to process your request at this time. This usually means that you are in violation of our API data access policy (https://crates.io/data-access). Please email help@crates.io and provide the request id ..."}]}
(some context https://matrix.to/#/!6oudZq5zJjAyrxL2uY:0upti.me/$Q25UIZLIjieJpO52C6VkOPoUYRBqtuEEN2QI4uUBIrM?via=nixos.org&via=matrix.org&via=node.marinchik.ink) | 09:37:02 |
