Should this be merged to Go core?

[nikolaydubina]

If this is really fast and without downsides, should this be in Go core?

Maybe folks in go-dev google group will get interest.

[shubhang93]

I don’t think so, although it looks fast on paper, one of the biggest advantages of go’s native channel is the ability to select across multiple channels and broadcast using the close function

[nikolaydubina]

Select, Close, Broadcast sounds like additional features, possibly implemented without deteriorating performance boost?

I am very curious what are the drawbacks of this approach. If there are no clear problems, this can be big thing like Ryu or more recent pdqsort.

I think we are standing on something big here.

[nikolaydubina]

cc @alphadose

[alphadose]

I have already opened an issue for it golang/go#52652

[alphadose]

We can gain even further optimizations by using runtime internal methods like gopark instead of spinning mechanism which is the current implementation thereby ensuring minimum required goroutine utilization without any performance drops

[MeteorsLiu]

Hope Go can really support this optimization. It's a certainly great project!
However, as far as i'm concerned, those guys in Google wonld't hear from people from the community most of time. That's a little bit tragedy. :(