attributes: support adding arbitrary fields to the instrument macro

[Kobzol]

This PR adds support for adding arbitrary key/value pairs to be used as fields to tracing::instrument.

Current syntax:

#[instrument(fields(key = "value", v = 1, b = true, empty))]

• Empty keys are supported
• If a key is not a single identifier, it's value is not a string/int/bool (or missing), is repeated or shares a name with a parameter, an error is reported

Remarks:

• I noticed that the Value trait is implemented only for integers, booleans, strings and Debug/Empty values. I wonder why floating point numbers are not supported?
• The testing infrastructure (MockSpan etc.) is awesome!
• Field keys in spans can be nested (for example parent.child = 1). We might achieve something similar here by transforming a path (parent::child) into a nested key (parent.child). However, I think that it would quite complicate the implementation (it's already pretty convoluted). And I'm not sure if we could quote! these transformed names properly, I had problems with quoting strings and had to use Ident for the keys (but I didn't spend too much time on this).

Fixes: #573

[hawkw]

This is great, thanks so much! I commented on a couple of minor nits, but it looks really good overall and I'll be very happy to merge it when the nits are addressed.

* Field keys in spans can be nested (for example `parent.child = 1`). We might achieve something similar here by transforming a path (`parent::child`) into a nested key (`parent.child`).

Ideally, I think it would be nice to be able to support nested field names here as well, and I'd really prefer to use the same syntax we do in the function-like macros (read: dotted names rather than paths). However, AFAICT, the only way to do this would be to stop parsing the macro input as NestedMeta, and write our own parser from scratch using syn's primitives...this might be a lot of work. So, for now, I'm happy to merge this without support for nested field names, as long as we open an issue to track adding that eventually with some details.

[hawkw]

Also: looks like CI is complaining about unnecessary parentheses; should be pretty easy to fix.

[Kobzol]

I changed the error messages and fixed clippy lint. After merge I'll add check for #562 in a new PR.

[hawkw]

Okay, this looks good, modulo one missing piece I forgot to mention earlier: can you add a brief note in the docs documenting the fields argument and how to use it (including the requirements that values be string, bool, or integer literals, and that dotted field names aren't currently supported)? An example or two in the docs would be great too, if you have the time.

Thanks!

[Kobzol]

Added docs and examples with basic usage. Let me know if that's good enough :-)

[hawkw]

this looks great; i had a couple last suggestions that you're free to disregard. i'll happily merge this when you feel like it's ready!

[Kobzol]

I hope that I understood your comments right :) Reformatted the documentation.

[hawkw]

Looks great! Thanks again!