What are the benefits of using `and` instead of `&&` when defining concepts?

Question

I found that some CppCon speakers used and instead of && to define concepts and used && in "normal" boolean expressions, but I can't figure out the benefits of doing so.

The only material about naming convensions of concepts that I can found is P1851: Guidelines For snake_case Concept Naming, but it says nothing about this.

Can anyone tell me why?

Example:

template <typename T>
concept boolean_testable
    = std::convertible_to<T, bool>
      and requires(T&& t) {
            { !std::forward<T>(t) } -> std::convertible_to<bool>;
          };

Instead of:

template <typename T>
concept boolean_testable
    = std::convertible_to<T, bool>
      && requires(T&& t) {
           { !std::forward<T>(t) } -> std::convertible_to<bool>;
         };

Although the complete substitutability between && and and caught my eye, I still can't figure out why some programmers deliberately choose to use and only when defining concepts. I noticed they still used && in other context (boolean expressions, T&&...). I can't figure out the benefits they got from such style.

I'm not asking for the difference between and and && but the motivation of using and instead of && only when in concept definition. There's no duplicate question as far as I know.

user17732522 · Answer 1 · 2023-01-29T11:23:58.387

9

There is absolutely no difference between && and and. As tokens they behave identically in all aspects except that the preprocessor # stringify operator respects their different spelling.

For example you can even replace T&& with T and and the code will still have identical meaning.

This is purely a style choice.

See [lex.digraph].

edited Jan 29 '23 at 11:23

answered Jan 29 '23 at 10:32

user17732522

53,019
2
56
105

really? Is this guaranteed or just an artifact of `T&&` being introduced only in C++11? I thought its only about the operator. – 463035818_is_not_an_ai Jan 29 '23 at 10:33
2

@463035818_is_not_a_number No, they are equivalent on the level of preprocessor tokens. I also fixed a minor mistake in my answer. The same goes for `T&` vs. `T bitand`. – user17732522 Jan 29 '23 at 10:35
@FeignClaims can you point to the relevant section in the standard ? I didnt bother to look. I only checked compilers and they do agree – 463035818_is_not_an_ai Jan 29 '23 at 10:38
@FeignClaims [`int and x = 42;`](https://godbolt.org/z/xbrzhdx9r) weird but compiles – 463035818_is_not_an_ai Jan 29 '23 at 10:39
@user17732522 My bad, wrong typo used when i try. – FeignClaims Jan 29 '23 at 10:40
I learned something new today: `T and` and `T bitand`. Two somethings actually! Along similar lines, my team at work has taken to doing `if (not ptr) ...` rather than `if (!ptr) ...` because we find it more legible. – Eljay Jan 29 '23 at 12:10

FeignClaims · Answer 2 · 2023-01-30T10:14:27.223

-1

In summary, the benefits might be:

and is an English word. Using and (along with or and not) makes the concept definition read like a beutiful verse.
In concept definition it's pretty common to have a lot of rvalue-reference, so using and instead of && helps readablity.

edited Jan 30 '23 at 10:14

answered Jan 29 '23 at 23:45

FeignClaims

193
8

What are the benefits of using `and` instead of `&&` when defining concepts?

2 Answers2