using std::swap; using std::begin; using std::end;

[LB]

http://ericniebler.com/2014/10/21/customization-point-design-in-c11-and-beyond/

For a long time I had never been concerned about having to write using std::swap; before calling swap() - I knew why I had to do it and even though I knew it was a nuance, I didn't care too much because it was trivial extra typing.

Then I saw this commit:
https://github.com/mosra/magnum/commit/65a2e5677a53602395e27d8fa20584dd7c851fbc

From what I understand, if the article's proposed changes are implemented into C++, then that commit wouldn't even have to exist.

ADL is a really powerful feature of C++ and I feel like it is held back by nuances like this. Do you think it's important that this issue be tackled as soon as C++17? Or do you think it can be set on the back burners for a while?

Personally, with the direction C++ is going, I think we will only need it to be fixed more and more, so for me, sooner is better.

[Cubbi]

Bjarne is fighting the hard fight at making f(x, y) and x.f(y) do the same thing in C++17 (or the closest thing possible to being the same), similar to how member and non-member operators are treated.. Microsoft is pushing for something similar, but giving precedence to x.f(y) (so that intellisense can work).. so things might change quite a bit,

[LB]

I've actually been thinking that myself, and was planning on making it a feature in my own language. It makes a lot of since and I would be very happy to see it come to C++.

[LB]

I just realized, if std::swap becomes an object with a templated operator()(), then code that explicitly uses template parameters will break. But, is this even an issue? Why would you ever write code such as std::swap<..., ...>(..., ...) in the first place?

[MiiNiPaa]

Why would you ever write code such as std::swap<..., ...>(..., ...) in the first place?

To use some questionable methods:

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

#include <iostream> #include <algorithm> struct foo { int i = 0; }; struct bar : foo { int j = 0; }; int main() { bar x, y; y.i = 3; y.j = 4; std::cout << x.i << ' ' << x.j << '\n' << y.i << ' ' << y.j << "\n\n"; std::swap<foo>(x, y); std::cout << x.i << ' ' << x.j << '\n' << y.i << ' ' << y.j; }

0 0 3 4 3 0 0 4

[LB]

https://ericniebler.github.io/std/wg21/D4381.html

[Little Bobby Tables]

my knowledge of UFI's stem from nim (http://nim-lang.org ), which doesnt have a this or a self, which is what allows them to work. how would UFI's work in c++ with a *this?

[LB]

Sorry, what does UFI stand for?

[Little Bobby Tables]

uniform function interface. its the name for x.foo(y) == foo(x, y); i was referring to:

cubbi wrote:

Bjarne is fighting the hard fight at making f(x, y) and x.f(y) do the same thing in C++17 (or the closest thing possible to being the same), similar to how member and non-member operators are treated.. Microsoft is pushing for something similar, but giving precedence to x.f(y) (so that intellisense can work).. so things might change quite a bit,

edit:
oops... im sorry guys... i meant uniform function call syntax (UFC syntax) not UFI.http://www.drdobbs.com/cpp/uniform-function-call-syntax/232700394

[LB]

Ah, I was more hoping to discuss customization points in this topic :)

[Little Bobby Tables]

fair enough. ill ask on IRC or perhaps in another thread