Stream Audio

[sigmaSd]

Hello @lemunozm,

I think you missed some conversation here #23 so I'm just opening a new issue.

I'm testing https://github.com/libcala/wavy crate to add this function, here is the current state https://github.com/lemunozm/termchat/compare/master...sigmaSd:audio_wavy?expand=1

This works but the major problem is that it currently spins the cpu a lot and slows down the entire program.

Maybe you can take a look at it and see if there is something to improve, or if this is a limitation of message-io maybe? since its flooded by audio message it can't keep the app responsive?

If its the latter case maybe message-io should offer some dedicated channels for huge data like streaming (since its the same problem with video but I guess streaming video sends less data?)

[lemunozm]

Hi @sigmaSd, Sorry for not answering, I currently have a lot of open fronts in mind, and because this part required somehow goes deep, I left to the next week.

By the way, I announced termchat in reddit and it has seemed to have a warm welcome, in part to your contributions with the file transfer and streaming. So thanks again for your work!

I like to be tracked this feature with a new issue! As soon as I'm free I get with it.

[sigmaSd]

Glad I helped!

No problem, I just reopened this issue so we remember this feature xD

Back to this issue, I prefer if I could use https://github.com/RustAudio/cpal just for the advertised android and windows support but for some reason I could not make it work on my pc, (no sounds from the speaker) that's why I used wavy here.

If I have more breakthroughs I'll post it here.

[lemunozm]

Hi @sigmaSd, I took a look at the changes. I like it a lot!

I will investigate the performance issue. Some days ago I was thinking about what you just comment about open a way to write/receive stream data without forcing it to make a message. I am working on some big refactorizations in message-io to offer this feature and others like adding new transports easily. I hope I could contribute from the basis to this performance increment.

Could you open a PR with this change to make better traceability in the comments/changes?