Ömer Sinan Ağacan
Ömer Sinan Ağacan
I'll write more later but just wanted to point out > Maybe we should have a write_ready that takes a buffer of TinyEv? We already have this, see `Conn::write_ready` (https://github.com/osa1/tiny/blob/master/src/conn.rs#L444-L455).
> I'm interested in bringing all the various messaging apps I use into one > terminal interface, where I can seamlessly switch between them. I currently > use Discord, Slack,...
As far as I understand you're thinking a futures/threads based API whereas I was thinking of a mio-based one. In mio-style code you can't tell if a socket your backend...
Btw is your code open source yet? I'd love to take a peek to get a better understanding at what we need from the backend API.
> I'm totally down for doing this with mio. I'm just not sure how mio figures out that my connection can be read from or written to... mio just decides...
Hmm I think this is not going to work -- for this API to work you'd need `read_ready` and `write_ready` methods for the entire stack. So, if you're using `slack`,...
I've been working on this and I started to think that we should implement a futures-based API and implement backends on top of that. I think for most backends we...
So I've been studying futures lately and I think I'm getting back to the thread-based API idea. futures ecosystem (tokio and co) are such a mess and it's impossible to...
I was excited about MAY too, but unfortunately it seems like it's deeply broken: https://github.com/Xudong-Huang/may/issues/6 the problem is not specific to MAY, we just can't implement work-stealing threads in Rust...
Thanks for reporting. I think showing the error just once in all tabs, and then showing the follow up errors only in the server tab makes sense. Showing the error...