On 4/28/24 12:31 PM, Peter Dimov via SG16 wrote:
> Tom Honermann wrote:
>> Is support for std::cout specifically required or would support for std::format()
>> and std::print() suffice? If std::cout support is specifically required, what
>> motivates that requirement?
> 30 years of "hello world".
>
> It's not 110% required, but it would be nice if it compiled and worked
> at least in the case where the string happens to be ASCII.
I certainly agree it would be nice.
>
>> During the meeting, we briefly discussed use of an opaque type for the return
>> type of name_of() and friends. I would like to see further exploration of this
>> idea prior to our next review. I'm envisioning a type something like the
>> following (This particular formulation follows existing precedent established by
>> the std::filesystem::path native format observers, [fs.path.native.obs]
>> <http://eel.is/c++draft/fs.path.native.obs> ).
>>
>>
>> class name {
>> std::string_view internal-representation; // exposition only.
>> name(/* unspecified */);
>> public:
>> constexpr std::string string() const; // ordinary literal encoding.
>> constexpr std::wstring wstring() const; // wide literal encoding.
>> constexpr std::u8string u8string() const; // UTF-8.
>> constexpr std::u16string u16string() const; // UTF-16.
>> constexpr std::u32string u32string() const; // UTF-32.
>> };
> I don't think this works for me. There's no way to get a constexpr
> char const* from this, and no way to get a constexpr string_view.

There is, but it doesn't have good ergonomics, probably doesn't work for
all use cases, and gcc/libstdc++ currently reject. You have to declare a
static constexpr buffer. Clang with libc++ and MSVC accept the following.

https://godbolt.org/z/5K8Pzo1Mr.

#include <string>
#include <string_view>
consteval std::string f() {
   return "test";
}
template<auto> void ft() {}
void g() {
   static constexpr std::string s = f();
   constexpr const char *p = s.data();
   ft<p>();
   constexpr std::string_view sv(s);
   ft<sv.data()>();
}

I make no claims that this is good enough. Perhaps others know how to
improve on it. Daveed seems to have some thoughts.

>
> (Unless we add a "reification" function that allows a consteval std::string
> to survive until runtime.)
>
> The current implementations allow
>
> constexpr char const* s1 = name_of(^float).data(); // NTCS
> constexpr std::string_view s2 = name_of(^float);
>
> The above does not.

Agreed.

Tom.

>
> Unrelated, I don't think we need to bother with char16_t and char32_t.
> One Unicode representation is fully enough.
>
>