Use universal references in append() method

[dota17]

This litte patch fixes issue #1046.

[BillyDonahue]

This is not a "universal reference" (which are now called "forwarding references").
It's an rvalue reference.

Anyway, for an "append" member function I'm looking at copying the API of standard container members like vector::push_back, which will always have a const T& and a T&& overload.

The Microsoft.co.il blog referred to in #1046 is having an internal problem right now so I can't read it, but I haven't heard of the {const T&, T&&} overload set as being a problem for append methods before.

I was able to find a mirror of it. ( https://dzone.com/articles/c-sink-parameter-passing ).

We're in the "write two overloads" camp currently, about which the author says:

""" The “only” problem with two overloads is that this approach doesn’t scale for multiple parameters."""

Well all we have are the ArrayIndex and the Value, and the ArrayIndex is an int.
The Value is the thing that needs to be treated specially, and that's what we're doing.
I think it's fine as-is, and we have a solution that doesn't require unnecessary moves of Value.

[BillyDonahue]

@dota17 can we close this PR now?

[dota17]

@dota17 can we close this PR now?

@BillyDonahue
if we close this PR, then how can we treat this issue #1046?
and how about the insert method?

[BillyDonahue]

@dota17 you are right. We should get a consensus on #1046 before closing this.

[dota17]

Closing. See the discussions in #1046.

[cdunn2001]

I hope Billy and Jordan don't feel as if they've wasted their own time on this. Personally, this was very informative for me, including the experiments. So I'm glad you tried this, Chen.

[dota17]

I hope Billy and Jordan don't feel as if they've wasted their own time on this. Personally, this was very informative for me, including the experiments. So I'm glad you tried this, Chen.

Yes, I agree, the experiments for this issue are very important. Thanks to Billy and Jordan's experiments. we have a clear and and reliable conclusions for that now.