implement support for ?Sync callbacks

[zetanumbers]

Resolves #263

[codecov]

Codecov Report

Merging #264 (5800990) into main (6acd555) will increase coverage by 0.50%.
The diff coverage is 83.33%.

❗ Current head 5800990 differs from pull request most recent head 2ffa136. Consider uploading reports for the commit 2ffa136 to get more accurate results

@@            Coverage Diff             @@
##             main     #264      +/-   ##
==========================================
+ Coverage   86.27%   86.77%   +0.50%     
==========================================
  Files          31       31              
  Lines        2478     2481       +3     
==========================================
+ Hits         2138     2153      +15     
+ Misses        340      328      -12     

Impacted Files	Coverage Δ
socketio/src/client/callback.rs	43.75% <ø> (+6.25%)	⬆️
socketio/src/client/client.rs	94.91% <ø> (ø)
socketio/src/client/raw_client.rs	77.63% <66.66%> (ø)
socketio/src/client/builder.rs	92.95% <100.00%> (ø)
engineio/src/asynchronous/client/async_client.rs	96.27% <0.00%> (+0.67%)	⬆️
engineio/src/asynchronous/client/builder.rs	91.93% <0.00%> (+0.80%)	⬆️
engineio/src/asynchronous/async_socket.rs	71.28% <0.00%> (+0.99%)	⬆️
...asynchronous/async_transports/websocket_general.rs	80.00% <0.00%> (+1.81%)	⬆️
...eio/src/asynchronous/async_transports/websocket.rs	91.89% <0.00%> (+2.70%)	⬆️
.../asynchronous/async_transports/websocket_secure.rs	90.78% <0.00%> (+2.78%)	⬆️
... and 2 more

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

[1c3t3a]

LGTM % comments. Thanks for the patch :)

[1c3t3a]

I am sorry, I missunderstood the mut thing here, could you drop commit 5800990? Then we just merge the first one.

[zetanumbers]

Why drop 5800990?

[1c3t3a]

Why drop 5800990?

Currently the builder method doesn't need the mut, but it would be possible (and tbh also expected) that the builder needs mut here. If we were not putting mut here, adding it would be a breaking change, so we might as well just have it there for the case that we ever need mut in a future version of the interface. I didn't realize that acquiring the lock and putting something there doesn't need mut as it uses interior mutability. So I thought we could just drop the allow(unused_mut) comment which was wrong.

[zetanumbers]

If we were not putting mut here, adding it would be a breaking change

It's nothing like that. This mut isn't even a part of the signature (check the docs.rs), but a part of the definition. That's because you can mutate immutable value by moving it out of the original immutable variable to the mutable one and mutate it however you want.

let a = "hello".to_string();
let mut b = a;
b.push_str(" world");

[1c3t3a]

Mhh fair point because we don't take &self here but a moved value. Then it make sense to remove it actually.