Add more Debug/Copy trait impls

[tmfink]

Also adds other useful impls.

[0xpr03]

Why do you need Hash ?

[0xpr03]

I don't see why everything has to be Hash and Copy, adding more compile time.
Copy makes everything unable to integrate anything that is clone, so any changes that require clone will be API breaking (especially config is such a candidate). And specifically for watchers it is not advised to implement Clone/Copy.

Also please don't format stuff unrelated to your PR, this makes it much harder to review.

[tmfink]

I don't see why everything has to be Hash and Copy, adding more compile time.

By making your types implement common traits (that make sense), it's easier to use the library. Due to orphan rules, it would be impossible for a consumer of the library to put the type in a HashMap or HashSet (without using a newtype wrapper and inventing a hash function). The C-COMMON-TRAITS API guideline has a good write-up.

Also, the difference compile-time is so small it's difficult to measure.

Copy makes everything unable to integrate anything that is clone, so any changes that require clone will be API breaking (especially config is such a candidate).

I'll restrict the Copy impls to the "C-like" enums (no associated data).
Tell me what you think it the next round of review. 😉

Also, it's already a breaking change to change variants of a public enum or to add variants to a non-non_exhausitive public enum (like Config).

And specifically for watchers it is not advised to implement Clone/Copy.

I'll remove the watcher Clone/Copy.
It's nice to get review from someone who knows the codebase better. 😄

Also please don't format stuff unrelated to your PR, this makes it much harder to review.

Sure--I kept them in separate commits, but I'll drop them.
The PR template is a little misleading:

Running `cargo fmt` and/or `cargo clippy` is NOT required but appreciated!

[0xpr03]

I'd remove the debug traits on internal stuff. The changelog addition is nice, though we haven't updated that in a longer time. I'm still fetching all changes manually per release and the "overall" thing will have to be regenerated anyway ;)

[0xpr03]

non-non_exhausitive

True, which is a reason we should probably make the fields non-pub and use a "builder" pattern.

it's easier to use the library

Yes and no. Not adding copy to bigger types prevents you from accidentally copying stuff instead when you'd probably rather use a reference.

[0xpr03]

Thanks for your patience and the PR! When CI finished I'll merge it.