Implement Hash for ParseError

[daxpedda]

This implements Hash for ParseError.
I assume that Clone and Copy wasn't added because potentially in the future some data might be added to ParseError?

I also replaced ParseError::_private with #[non_exhaustive], kinda the new Rusty way to do this. Let me know if you want me to split this into a separate PR.

[madsmtm]

I'm not sure #[non_exhaustive] really works on empty structs? See this playground

[notgull]

I also replaced ParseError::_private with #[non_exhaustive], kinda the new Rusty way to do this. Let me know if you want me to split this into a separate PR.

The docs render differently and show an empty struct rather than a private struct. So I would prefer not to do this.

[daxpedda]

I'm not sure #[non_exhaustive] really works on empty structs? See this playground

It does work, just not on a intra-crate level, but on a inter-crate level.
You can try it in a local example.

I also replaced ParseError::_private with #[non_exhaustive], kinda the new Rusty way to do this. Let me know if you want me to split this into a separate PR.

The docs render differently and show an empty struct rather than a private struct. So I would prefer not to do this.

Looks like this:

I agree that's worse.

[daxpedda]

Never seen this before:

cursor-icon/src/lib.rs

Line 12 in 9a8f7d0

#![cfg_attr(feature = "cargo-clippy", deny(warnings))]

What is this supposed to do and how exactly should this be fixed?

[kchibisov]

@daxpedda the point was to not allow building of the struct while remaining it public, probably it does achieve that, but I don't really see a value in changing the _private? Like it adds nothing really.

[daxpedda]

@daxpedda the point was to not allow building of the struct while remaining it public, probably it does achieve that, but I don't really see a value in changing the _private? Like it adds nothing really.

I already reverted that because of @notgull's feedback.

What it would have done is to use the recommended way Rust achieves exactly this, instead of having to add a field that does nothing, which is basically a workaround. Most people consider adhering to common code practices a code quality improvement.