Prefix the name with what it’s for. For example, I’ve previously got a SoundFontError from opening soundfont file.

“Error” is already used by std::error::Error. It might not be imported by the code that imports your error type, but I think it’s better to give it distinct name.

The other thing is that you might want to use more than one library. Typical imports at the top of the file might look like this:

use bingbong::{BingBong, BingBongError, SomethingElse};
use bogos::binted::crotchidizer::{Crotchidizer, CrotchidizerError};

If both libraries named their error enums just “Error”, the collision could be worked around, but it’s an unnecessary extra step:

// Not sure how renaming affects compiler hints.
use bingbong::{BingBong, Error as BingBongError, SomethingElse};
use bogos::binted::crotchidizer::{Crotchidizer, Error as CrotchidizerError};

or if you want to avoid renaming:

use bingbong::{BingBong, SomethingElse};
use bogos::binted::crotchidizer::{self, Crotchidizer};

/* ... */

    match result {
        Ok(value) => return value,
        Err(bingbong::Error::MissionFailed) => panic!(),
        Err(bingbong::Error::UrMom) => todo!(),
        _ => unreachable!(),
    }
    
    if let Err(crotchidizer::Error::SomethingsWrong) = result2 {
        // ...
    }
    

If the screenshot had followed conventions, the message would say something like this:

could not convert error type `BingBongError` to `MyAppError`

You really should avoid naming your type plain “Error”

I don’t know what app you’re using, but that spoiler tag isn’t part of the spec.

The sidebar’s gone because I removed it :)

I’ll try to incorporate that info elsewhere. At this stage, a lot can change on a whim.

Yeah, I’d love to have some kind of full SSR solution that doesn’t require scripts at user’s end, but it’s probably not possible with this.

I wish there were more people building frontends. Since Lemmy is properly decentralized, there’s a lot of potential for interesting instance diversity.