Juhana Helovuo

I cannot remotely tell you what is going wrong here, but here are some debugging instructions: 1. In your Rust application enable logging, as done [here](https://github.com/jhelovuo/RustDDS/blob/master/examples/shapes_demo/main.rs#L315) and [here](https://github.com/jhelovuo/RustDDS/blob/master/examples/shapes_demo/logging-config.yaml). The root...

> the QOS 'history depth' property is stored in most places as an `Option`. It seems that usually `Some(..)` means a finite depth, and `None` means 'keep all'. It's clear...

That's right. Provided that you take care that history depth is always serialized as 32-bit integer, even when `usize` is 64-bit, then the rest of the changes should be quite...

The original design decision to use `mio` was made some years ago (2019?) when asyncs were not yet up to the task. Are you suggesting that the application-facing API would...

Yes, it would be an advantage to have an async application API, so RustDDS could be used by async applications directly. But moving the implementation itself on top of async...

Does "spawn tasks" mean here approximately the same as "start background threads", except using the async mechanisms instead of OS threading? I think that is precisely what we would need,...

Technically, this crate is a fork of rtps-rs. Practically, this crate is a working DDS/RTPS implementation, whereas rtps-rs is an work-in-progress RTPS implementation and no DDS layer.

> i've been looking at this a little further (and let me preface this by first saying i'm emphatically _not_ an expert in low-level network programming!). I think this crate...

Sounds like a lot of work, but a chance to learn a lot on the way. Just check that you have the latest specs, i.e. DDS 1.4 and RTPS 2.5....

I have an alternative idea, although it needs some refinement: We could keep the current dp_event_loop and discovery event loop, but provide an alternative, async API so that RustDDS could...