Bela Schaum

Nice workaround! > do not tell anyone This issue is linked already to the llvm-project bug.

The switch is not required to handle all cases (especially unrecognized values), so when explicitly isn't set the result (or the type) for the `magic_enum::enum_switch` call, the result type cannot...

If I wrote a swich, I don't need to write all case / default. I think this is an another use-case, and probably not a `switch` but a `visit`. I...

Hi! The problem here is not the magic_enum library, but that MSVC handles the nested lambdas. As you [can see](https://godbolt.org/z/qWKvzMcWT), if the nested `enum_switch` not use the outside type, the...

From the library side: This issue can be eliminated with the following techniques: - make `magic_enum::enum_for_each` and `magic_enum::enum_switch` to handle variadic arguments, and generate all enum combination calls, or the...

Probably it would be nicer if I separate `test.cpp` testing parts into `tests.h`, because I used an ugly `cpp include` hack for testing.

Without `MAGIC_ENUM_OPT_ENABLE_NONASCII` it can be easily support wide char, because (usually) wchar_t encoding backward compatible with ASCII, and transition can be made compile time. With NONASCII characters the conversion is...

Thanks @alexkaratarakis , but I have only C++17 std available compiler, fixed-containers use C++20

I think these containers not fits to the `magic_enum.hpp` header file. Probably it need to put to an another header file, like `magic_enum_containers.hpp`. But then the library looses the "only...

Something like that, but with constexpr functions, and stl-like api (in C++17).