Specialize the conversion between wchar_t* and Go string on Windows only

[koron]

This will fix #6

[sstallion]

@koron Thank you for your contribution! Could you provide a little more detail on the issue you're seeing? Changes look pretty good, though I do worry a little about some of the casting being done (namely interchanging wchar_t * and *uint16, which can have different alignment). This might need some deeper changes to support native UTF-16 on Windows.

[sstallion]

Thinking about this a little more, what would you think of just changing references to *C.wchar_t to unsafe.Pointer? We can keep the original names but that may be the safest way to avoid making assumptions about alignment.

[koron]

Could you provide a little more detail on the issue you're seeing?

The root cause of the problem is C.wcstombs() and C.mbstowcs can't treat Chinese/Japanese/Korean in Windows.
It is hard to describe why those functions doesn't work,
it's a complex factors (code page, encoding, c-runtime, compiler used by CGO...) involved to dealing CGO in Windows.

namely interchanging wchar_t * and *uint16, which can have different alignment

And sizeof(wchar_t) is 2 in Windows. So those alignments are same between wchar_t and uint16.

See also: https://learn.microsoft.com/en-us/cpp/cpp/char-wchar-t-char16-t-char32-t?view=msvc-170

The wchar_t type is an implementation-defined wide character type. In the Microsoft compiler, it represents a 16-bit wide character used to store Unicode encoded as UTF-16LE, the native character type on Windows operating systems.

Anyway, in Windows, what is needed is to convert between UTF-16 (wchar_t[]) and UTF-8 (Go string).
It is defined in syscall package for Windows or x/sys/windows package.
I thought it was a bit confusing that the syscall package had different contents for each OS, so I used the x/sys/windows package.

what would you think of just changing references to *C.wchar_t to unsafe.Pointer?

I guess the cast to unsafe.Pointer wasn't needed, I've removed it from the PR.

[sstallion]

The root cause of the problem is C.wcstombs() and C.mbstowcs can't treat Chinese/Japanese/Korean in Windows.
It is hard to describe why those functions doesn't work,
it's a complex factors (code page, encoding, c-runtime, compiler used by CGO...) involved to dealing CGO in Windows.

Gotcha. If you could add a comment to this effect and include a copyright header at the top, we should be good to go. Thanks again!

[koron]

I added comments and update the PR.
Please check it.

Thanks.

[sstallion]

Thanks @koron! I'll get this merged ASAP.