Remove unneeded Debug trait requirement from BasicSnippet and DeprecatedSnippet

[Sword-Smith]

No description provided.

[jan-ferdinand]

Does this trait requirement interfere with anything? Do the #[derive(Debug)] interfere with anything? It is common practice to #[derive(Debug)] wherever possible, unless there's a good reason not to do so, of which I can find few and none that seem to apply here.

[Sword-Smith]

Does this trait requirement interfere with anything? Do the #[derive(Debug)] interfere with anything? It is common practice to #[derive(Debug)] wherever possible, unless there's a good reason not to do so, of which I can find few and none that seem to apply here.

I see two arguments for removing Debug

1. It puts a Debug trait constraint on fields in snippet structs that implement BasicSnippet
2. It's not needed anywhere as a snippet should be fully identified by its entrypoint. So adding it gives the compiler more work than necessary.

[jan-ferdinand]

1. I return to my first question; does the trait bound interfere with anything?
2. I'd be very surprised if that additional workload was significant.

In the end, this is your codebase; do as you see fit.

[Sword-Smith]

1. I return to my first question; does the trait bound interfere with anything?

   2. I'd be very surprised if that additional workload was significant.

In the end, this is your codebase; do as you see fit.

1. Not currently. I could, theoretically, in the future.

[Sword-Smith]

I guess my opinion is more based on aesthetics and how I think you should inspect values of this type. Aesthetically I don't like to impose trait bounds that are not needed, and I think that you should inspect the the type of a snippet through its entrypoint name, not through its debug output.

[jan-ferdinand]

If that is the case, then how about the following?

impl Debug for dyn BasicSnippet {
    fn fmt(&self, f: &mut Formatter<'_>) -> std::fmt::Result {
        write!(f, "{}", self.entrypoint())
    }
}

This would also allow the #[derive(Debug)] on the various types to stay in place.

[Sword-Smith]

Your suggestion works. But I should also like to point out that no Debug derivations are removed with this PR, it's only the requirement that is removed.

[jan-ferdinand]

What about removal of line 35 in all.rs, line 31 in filter.rs, line 83 in inner_function.rs, and line 37 of map.rs?

[Sword-Smith]

What about removal of line 35 in all.rs, line 31 in filter.rs, line 83 in inner_function.rs, and line 37 of map.rs?

You're right. It got removed from higher-order snippets.