about use std::sync::Mutex

[yaobo-lab]

What happened?

the project ManagedState struct and SharedServiceDaemon : use std::sync::Mutex;

but when tauri command async use tauri::async_runtime::spawn; like this

the standard library Mutex lock in the tokio runtime result in poor performance? use tokio::sync::Mutex; Replace it?

Steps to Reproduce

about use std::sync::Mutex

Operating System

Windows

Install Method

None

Relevant log output

[hrzlgnm]

But your example does not use that mutex at all as poll_can_browse does not access queriers not the daemon of the state

[hrzlgnm]

And otherwise where we use the mutex we clone the daemon before moving it into the async task

[yaobo-lab]

But your example does not use that mutex at all as poll_can_browse does not access queriers not the daemon of the state

I'm just giving an example, other commands are use

[hrzlgnm]

But not in async code at all so replacement does not make any sense to me.

[yaobo-lab]

And otherwise where we use the mutex we clone the daemon before moving it into the async task

the std mutex lock os threads 。But tokio::sync::Mutex not ？

[hrzlgnm]

You mean std::sync::Mutex blocks threads? But there is no issue with that as those locks are pretty short in our case, those are just released again shortly after moving on to async code.

[yaobo-lab]

You mean std::sync::Mutex blocks threads? But there is no issue with that as those locks are pretty short in our case, those are just released again shortly after moving on to async code.

You can search the difference between tokio:: sync:: Mutex and std:: sync:: Mutex。（tauri async is use tokio runtime by default） When concurrency is high, you will encounter

[yaobo-lab]

In fact, both can be used, just in terms of performance. That's more suitable

[hrzlgnm]

In fact, both can be used, just in terms of performance. That's more suitable
You can search the difference between tokio:: sync:: Mutex and std:: sync:: Mutex。（tauri async is use tokio runtime by default） When concurrency is high, you will encounter

I've in fact searched for the difference, and from my point of view, it would be wrong to switch to tokio::sync::Mutex, as we never use a mutex in a async context, nor do we create many threads. Therefore the argument with encountering performance impacts in high concurrency situations is just wrong.