2024 edition update and other minor refactors#199
Conversation
|
Seems like a spurious CI failure(?) |
|
Switched to a single threaded runtime for tokio, shouldn't have any impact but just minor efficiency gains |
Please don't, AFAIK single-threaded runtimes can quite easily deadlock when not being careful about it, and we're not performance critical anyways so the overhead should be negligible |
I could only really imagine that stdout/stderr locking in the updating/verifying code would be a cause for worry here, which would hopefully be caught easily in testing before it could become a issue . Even if we don't switch to single threading I would prefer to have only 2 threads for the sake of not being excessive. |
|
Since it's a relatively minor change code wise anyhow, I'll spin the single threaded tokio runtime off into another PR, I'll keep testing it in the meantime to see if I get any weird edge cases with my use. |
Coca162
force-pushed
the
minor-refactors
branch
2 times, most recently
from
daf7155 to
51c3829
Compare
|
Added back the comment about clap's version that was accidentally stripped out with #186, I assume it still applies so |
Coca162
force-pushed
the
minor-refactors
branch
3 times, most recently
from
9be3b49 to
0539b0d
Compare
|
Not sure about why the CI is failing on snix now, locally I can run it fine with this hash... |
IIRC snix has some store paths in their files and those will be reference scanned and then nix complains about undeclared / impure references. I'm not sure why that doesn't surface for you locally, but I can reproduce it on my machines. Perhaps it is best to just delay the snix package upgrade, if we can. |
Coca162
force-pushed
the
minor-refactors
branch
2 times, most recently
from
9c14367 to
98988bf
Compare
|
Pinned it to the revision we are using already in the manifest |
- Using edition 2024 for all crates - Fixed all checks/clippy warnings - [`clippy::wildcard_imports`](https://rust-lang.github.io/rust-clippy/stable/index.html#wildcard_imports) now warns for libnpins - `anyhow::Result` is now always fully qualified in libnpins - [Formatting code with the new edition](https://doc.rust-lang.org/edition-guide/rust-2024/rustfmt-formatting-fixes.html) - `cargo update` - Turn off default features for `nix-compat` and set exact revision
Coca162
force-pushed
the
minor-refactors
branch
from
98988bf to
7921415
Compare
3 participants
Some cleanup work for the codebase:
clippy::wildcard_importsnow warns for libnpinsanyhow::Resultis now always fully qualified in libnpinscargo updatenix-compatand set exact revision