`<xloctime>`: `time_put_byname` doesn't encode the output in UTF-8 even if the locale name contains ".UTF-8"

[cpplearner]

Describe the bug

time_put_byname seems to use the system code page even if the locale name contains .UTF-8.

For example, in the ja_JP.UTF-8 locale, "土曜日" (which means "Saturday") is encoded as "\xCD\xC1\xEA\xD7\xC8\xD5" on my computer, but it's UTF-8 representation is "\xE5\x9C\x9F\xE6\x9B\x9C\xE6\x97\xA5".

The libcxx test std/localization/locale.categories/category.time/locale.time.put.byname/put1.pass.cpp expects the output to be encoded in UTF-8 in this case.

time_put_byname is super old and rarely used, and fixing this would be a behavior change. I don't know if it's worth the trouble.

Command-line test case

D:\test>type test-time-put-byname.cpp

#include <cassert>
#include <ctime>
#include <iostream>
#include <locale>

class my_facet : public std::time_put_byname<char, char*>
{
public:
    explicit my_facet(const std::string& nm, std::size_t refs = 0)
        : time_put_byname(nm, refs) {}
};

int main() {
    const my_facet f("ja_JP.UTF-8", 1);
    std::tm t = {};
    t.tm_wday = 6;
    std::ios str(nullptr);

    {
        char buf[200];
        char* iter = f.put(buf, str, '*', &t, 'A');
        std::cout.setf(std::ios::hex, std::ios::basefield);
        for (char* p = buf; p != iter; ++p) {
            std::cout << (int)(unsigned char)*p << ' ';
        }
        std::cout << '\n';
    }
}

D:\test>cl /EHsc test-time-put-byname.cpp

Microsoft (R) C/C++ Optimizing Compiler Version 19.50.35725 for x64
Copyright (C) Microsoft Corporation.  All rights reserved.

test-time-put-byname.cpp
Microsoft (R) Incremental Linker Version 14.50.35725.0
Copyright (C) Microsoft Corporation.  All rights reserved.

/out:test-time-put-byname.exe
test-time-put-byname.obj

D:\test>test-time-put-byname.exe
cd c1 ea d7 c8 d5

Expected behavior

The output is e5 9c 9f e6 9b 9c e6 97 a5.

STL version

117ca96

Additional context