The playground link doesn't work, and I'm not sure what you are suggesting.

Github does not add the _ at the end of the play link to the actual link:
https://play.golang.org/p/4R-IQsF3p_

Sorry, terrible playground I created over there. Did a correct one over here:
https://play.golang.org/p/Hu3qX6aSr-

Basically I am returning a nil error that was already checked and it's known that it's nil. I think vet will be able to catch this but I am not sure.

Thanks

Thanks. I think that you are asking for vet to look for cases where we return a value of type error but where vet can prove that the value is always nil. For vet tests we look for three characteristics; see cmd/vet/README. This seems correct, and I imagine that we can make it precise, but how frequent is it?

Yes, exactly. I think it's rare as well but when it happens its pretty hard to see it in the middle of all that code.

Perhaps it could also be used for other types that aren't errors - making it way more useful but then it might have a performance penalty.

By the way, thank you for that cmd/vet/README, hadn't seen it before :)

Perhaps this can be closed as it doesn't match the "frequency" rule.

Thank you

What about:

val, err := ret()
if err != nil {
	panic("DEBUG")
}
return val, err

where the panic is just for debugging? Does vet now complain that err is statically nil in the return? It seems like a very specific check. So in addition to frequency above I'd raise some questions about precision. But it seems like the consensus is already that it's too special, so closing.