spec: apparent contradiction for passing of slice argument to variadic parameter

[myitcv]

Version of November 16, 2018

Section "Passing arguments to ... parameters" says:

If the final argument is assignable to a slice type []T, it may be passed unchanged as the value for a ...T parameter if the argument is followed by .... In this case no new slice is created.

Notice the use of the word "may". This appears to contradict what I understand to be the case in practice (i.e. that it will be passed unchanged, for reasons of efficiency) but also the example that follows:

... who will have the same value as s with the same underlying array.

Notice the use of the word "will".

cc @griesemer

FYI @mvdan @rogpeppe

[ianlancetaylor]

In this case "may" refers to the programmer, and "will" refers to the implementation. So I think it is correct.

But I'll let @griesemer decide whether to change the wording to avoid confusion.

[mvdan]

If it serves as a data point, all three of us (Roger, Paul, and myself) found the wording a bit confusing :)

[myitcv]

Thanks @ianlancetaylor. So the key part of that first paragraph is actually the final sentence:

In this case no new slice is created.

[griesemer]

I suggest s/may be/is/ in that sentence. That is, it would say:

If the final argument is assignable to a slice type []T, it is passed unchanged as the value for a ...T parameter if the argument is followed by .... In this case no new slice is created.

Does that seem clearer?

[myitcv]

@griesemer yes that seems much clearer (to us three at least!). Thank you

[gopherbot]

Change https://golang.org/cl/167380 mentions this issue: spec: clarify wording on passing slice arguments to variadic functions