Roland Fredenhagen
Roland Fredenhagen
@udoprog what is your opinino on this? > One thing to decide is whether to require the `Self::` as protocols are likely often associated functions. > > This would mean...
> Seems fine. For the way it's implemented you should even be able to omit `self::` if the function is unambiguous. currently yes the question was about changing this, i.e....
> PR looks to be in pretty good shape with a few things that can be done later. Would you be OK with me going over, clean a few things...
I would be willing to implement this.
Access to doc comments would be the next helpful thing, I'm not pressed on having it ASAP as I'm still very much in the beginning of my scripting tool and...
```rs fn error() -> Result { let mut module = Module::with_crate_item("std", ["error"]); fn display_error(error: &rune::Error, buf: &mut String) -> fmt::Result { write!(buf, "{error:?}") } module.ty::()?; module.associated_function(Protocol::STRING_DISPLAY, display_error)?; Ok(module) } ```
it is the current location actually https://github.com/rune-rs/rune/blob/e5b7a3a1095768b26fa05ec4e2857aba78a019f7/crates/rune/src/any.rs#L86
It would be neat though to have something to easily produce readable errors, and anyhow is quite simple to use :D. Another thing is that the hash is really not...
> What anyhow gives you is stacktraces and the ability to wrap any error. If those added capabilities aren't use, they don't need to be their own type. Yes, for...
I will try next week when I have access to the old windows laptop of mine, but sounds promising.