-
Avoid use of `anyhow` in Regorus.
Turns out `ThisError` is not `no_std` compatible.
----------------------------------------------------------------------------------
Original description:
…
-
Right now the codebase are full of unwrap of option and error. This will cause Rust to panic which is an undesirable behaviour. In Rust, panics are generally reserved for unrecoverable error. A sugges…
-
Adding explicit errors is helpful for library consumers, `anyhow::Error` doesn't really provide any indication on a programmatic level of WHY it failed, so it becomes impossible to "catch" a specific…
-
```rust
let mut criu = rust_criu::Criu::new().expect("unable to new a criu");
match criu.get_criu_version() {
Ok(version) => info!("criu version: {}", version),
Err(e) => erro…
-
Since this application does not have a GUI at the moment but has a comparatively wide target audience, some clearer error message such as those provided by [anyhow](https://crates.io/crates/anyhow) in…
-
Eyre is a close clone of anyhow which aims to provide extra configurabilities and more dynamic features compared to anyhow.
The library has been able to achieve its purpose and provide more elabora…
-
I see that `cameleon` uses `anyhow::Error`, but very lightly: https://github.com/search?q=repo%3Acameleon-rs%2Fcameleon%20anyhow&type=code (only for 2 `Io` error variants).
Anyhow specifically isn'…
-
We have been too liberal of our use of `anyhow`. For any `pub` function in our libraries, we should be using `thiserror` instead of `anyhow`, to get more type safety and explicitness of what can go wr…
emilk updated
3 months ago
-
This kinda piggy-backs on to trait support. In general, it seems that when a trait is designed to be implemented by multiple different crates, for any `Result`, E is either a generic type associated w…
-
To ensure consistency and provide more meaningful failure messages during testing, we should replace all instances of `unwrap()` in tests with proper error handling using `anyhow::Context`. This will …