P/Invoke and UTF-32

[iahung2]

On Linux, wchar_t is UTF-32. CharSet.Unicode is UTF-16.

[huoyaoyuan]

A UTF-32 char is just an uint. You can also use the more strongly-typed System.Text.Rune.

[colejohnson66]

What is your question? Is it how to marshal as UTF-32 in a LibraryImport?

[iahung2]

Yes. My own research shows it's currently impossible. So, I have to come up with workarounds.

[tannergooding]

On Linux, wchar_t is UTF-32.

On Windows wchar_t is notably UTF-16. It's one of the places where different platforms have different definitions for basic C/C++ types.

CharSet.Unicode is UTF-16.

Yes, which matches the .NET char type. You would need to explicitly use int or uint (preferably matching the signedness of wchar_t for the underlying platform) when targeting Unix based platforms where wchar_t is 32-bits.