Hacker News Reader

I think that's too strong.

You could have two executors with no problem.

What might cause issues is if code running on one executor needed close interaction with the other - such as running a task on the other executor.

Most of the time this is just avoided by using only one executor. But that's not the only way.

View on HN

That's a good clarification. What matters is if tasks "running on" one executor end up spawning / blocking on a task "on another executor". You can definitely have "side by side, who cares". The danger though is that you're just using some library, and if it suddenly assumes it can do tokio::task::block_in_place but you were using some other executor you get "Hey! No runtime!".

Reply on news.ycombinator.com