feat: expose the returned remote capabilities on the Client

[RomFouq]

This would tackle #195 by storing the remote capabilities (obtained with the new session response) in the Client, and exposing them via a public getter for later use.
This internally leverages the already existing webdriver::response::NewSessionResponse to map the response.

Being able to access the remote capabilities is necessary to be able to obtain the webSocketUrl remote capability and establish a WebDriver BiDi connection.

Notes:

• The remote capabilities are stored in and exposed on the Client because it was the easiest, not sure if it would make more sense to have them on the Session instead.
• The public getter is named .remote_capabilities(), as it exposes the capabilities returned by the remote end, maybe you would prefer another name as I don't think fantoccini uses the local end/remote end terminology from the W3C specification.
• I've replaced the .remove()/.insert()-back-in-case-of-error JSON value extraction with .get()/.to_owned() as we now also need to access capabilities as well; do you think it would be better to avoid this clone?

[jonhoo]

This is a great addition, thank you! Left a few questions, but none of them major.

[RomFouq]

I updated the PR and addressed your comments:

• The whole NewSessionResponse is now stored in the Client: as
  NewSessionResponse does not implemented Clone and we want Client to be
  Clone and still Sync, I put it into an Arc (and an Option, since there
  is no obvious default NewSessionResponse).
  That response is exposed with the getter session_creation_response() (I
  didn't name it new_session_response() because this looks too much like a
  constructor function to my taste).
• The remote capabilities getter method is now simply
  named capabilities().

[codecov]

Codecov Report

Merging #223 (f1256f5) into main (9960984) will increase coverage by 0.27%.
The diff coverage is 87.50%.

Files Changed	Coverage
src/wd.rs	76.92%
src/session.rs	92.30%
src/client.rs	100.00%

[RomFouq]

Oh right, fixed that by introducing our own NewSessionResponse, marking it non_exhaustive as you said, making in Clone, and deriving Serialize and Deserialize.

[jonhoo]

Nice! Very nearly there 🎉

[jonhoo]

Also, you may want to merge from main to get a bunch of test fixes.

[jonhoo]

Thanks!

[RomFouq]

The test checking that fantoccini can compile with the minimum version of each dependency seems to be failing because we try to derive Eq on NewSessionResponse, which contains Capabilities, a serde_json::Map, while serde_json has only been implementing Eq on Map since version 1.0.50, and current minimum version in fantoccini is 1.0.25.
Can we bump serde_json in this PR, or should I just not derive Eq on NewSessionResponse for now?

The other failing test seems unrelated.

[jonhoo]

Totally fine to bump the minimum version of serde_json!

[RomFouq]

Great! So I bumped serde_json to the newest version (released yesterday). It shouldn't increase the MSRV.

[jonhoo]

I'd generally prefer only bumping it to what we actually need (what goes in Cargo.toml is a minimum version after all). But not the end of the world if it preserves MSRV

[jonhoo]

Thank you!

[jonhoo]

Released in v0.20.0-rc.7 🎉

[RomFouq]

Thanks a lot for your help on this PR and for the release!