Clang-Tidy warnings for static coroutine methods initial_suspend and final_suspend

Question

Can (and maybe should) my coroutine promise class MyCoroutine::Promise have static or non-static initial_suspend and final_suspend methods?

static std::suspend_always initial_suspend() noexcept
{
    return {};
}

static std::suspend_always final_suspend() noexcept
{
    return {};
}

While this works fine (tested with GCC 13), Clang-Tidy complaints "Static member accessed through instance", when using this coroutine type:

MyCoroutine foo() // Clang-Tidy warning here
{
    co_return;
}

This makes sense because "when a coroutine begins execution, it [...] calls promise.initial_suspend()" (see https://en.cppreference.com/w/cpp/language/coroutines.html) via instance, not via promise type.

So should I remove the static keyword from these methods (and make them const)? A this pointer is not required, but probably the compiler can inline and optimize these calls, anyway. Or are static methods a good idea here and is there a better way to deal with this warning?

Some of the examples from https://en.cppreference.com/w/cpp/language/coroutines.html do have static initial_suspend and final_suspend methods, some don't.

That's just a bug in clang-tidy. The implicit call to initial_suspend inserted by the compiler should not be triggering style criticisms like that. — Brian Bi
– Brian Bi, Commented Aug 25 at 22:11

Nicol Bolas · Accepted Answer · 2025-08-25 15:20:06Z

4

Making them static would change nothing meaningful about the program. Non-static, non-virtual members don't have any inherent limitations to compiler-based optimizations compared to static ones. If the function definitions are in the header/module interface, then the compiler can inline them or otherwise do definition-based optimizations regardless of whether they're static or not.

answered Aug 25 at 15:20

Nicol Bolas

481k66 gold badges861 silver badges1.1k bronze badges

Sign up to request clarification or add additional context in comments.

3 Comments

Jens Aug 26 at 8:12

Thank you, Nicol. So you would prefer non-static methods, while Brian seems to prefer static ones. I guess both is alright.

Nicol Bolas Aug 26 at 13:47

I thought he just said that it shouldn't be a warning. That's not expressing that he "prefers static ones". Personally, I don't see the point of the "static member accessed through instance" warning in any cases.

Jens Aug 27 at 8:31

You are right. And I tend to make these methods non-static now, although Clang-Tidy might give another warning (see clang.llvm.org/extra/clang-tidy/checks/readability/…, I haven't tested it in this case, though). While the this pointer is not used (at least not now), logically these methods belong to the instance, not the type.

Collectives™ on Stack Overflow

Clang-Tidy warnings for static coroutine methods initial_suspend and final_suspend

1 Answer 1

3 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

3 Comments

Your Answer

Sign up or log in

Post as a guest

Related