Paul Fultz II
Paul Fultz II
> Yes, it needs to be a compiler that supports HIP, which right now is either amdclang, hipcc or cray cpe. Not just amdclang, but any recent versions of upstream...
> clang currently can set the proper include path by itself for HIP. You mean clang will set the include path for the private headers it uses? If so, then...
Having a framework to build adaptors can simplify the building of adaptors in the library, and simplify how they could be defined by the user.
> You are right, but people want to know how to define flat adaptors as the ones you have defined. But there is a lot detail things that goes on...
Builders could be provided to build adaptors. Here are the different builders I am thinking about: - `unary_adaptor_builder` - This builds an adaptor taking one function - `binary_adaptor_builder` - This...
The syntax is the same for constexpr. Also, for none of the adaptors I have a constexpr example. This is partly because a few platforms have limited support for constexpr(such...
Interesting, I haven't heard of `decay_copy` before. It looks almost the same, except `fit::decay` also unwraps references.
> template cannot be MoveConstructible Well, the template itself cannot, but after the types are placed in the template, it must be MoveConstructible. > It is not clear from the...
> Yes, but this mean one template for each E in expected. No, template aliases can be defined at class scope.
Something like this: ``` cpp template struct e { template using apply = expected; }; auto x = make(1); ``` Although it may not be quite as slick as the...