Bastian Kersting
Bastian Kersting
This is closed in #229 isn't it?
Yeah makes sense! I appreciate every contribution :) You can basically fully lean on the current implementation and try to retrofit the changes made to the protocol. A feature flag...
> @1c3t3a Compiler does not complain until I create a trait object `Box`: > > --> socketio/src/client/reconnect.rs:40:8 | 14 | pub trait ClientTrait { | ----------- this trait cannot be...
Sounds good! From my experience it was never bad advice to listen to @nshaaban-cPacket :) I would accept any PR that fixes the existing questions.
We might also add a coverage checker to the repo, maybe [Codecov](https://github.com/marketplace/codecov/). But this only makes sense if the containerized tests are up and running.
I've thought this through and IMO the tests in each file are clearly easier to grasp and understand than when they're in a separate file.
Further information about how the pipeline and the test suite is set up can be found at `ci/README.md`.
Current open points: - fix tests so that it is ensured that callbacks are called correctly - ~~investigate UB that might arise when cloning the client~~ - ~~put async callbacks...
@nshaaban-cPacket What alternatives to async callbacks would you consider? A possible alternative would be to somehow use the implementation of `Stream` to react to a packet and provide means to...
I added the `async-callbacks` feature flag. As there is currently no other alternative to async callbacks, I added it as a default feature when `async` is enabled. Anyhow the `ClientBuilder`'s...