From: Sam Smoot <ssmoot@...>
Date: 2006-10-10T02:00:15+09:00
Subject: Should this work? C++/CLI and Ruby UTF-8 Strings

I need to convert a .NET Unicode String to a UTF-8 byte-array to create
a Ruby String that I can display on a UTF-8 encoded web-page.

I don't know C++ really, but a friend and I have stumbled our way
through most of what we need. I need to figure out this String->UTF-8
stuff though.

Here's the code:

  VALUE ToRubyString(String^ string) {
    if (string == nullptr) return Qnil;
    IntPtr ptr;
    try {
      char *ret =
(char*)malloc(4+System::Text::Encoding::UTF8->GetByteCount(string));
      cli::array<unsigned char>^ char_array =
System::Text::Encoding::UTF8->GetBytes(string);
      cli::pin_ptr<unsigned char> char_pin_ptr = &char_array[0];
      strcpy(ret, (char*)char_pin_ptr);
      return rb_str_new2((char*)(void*)ret);
      // ptr =
System::Runtime::InteropServices::Marshal::StringToHGlobalAnsi(string);
      // return rb_str_new2((char*)(void*)ptr);
    } finally {
      System::Runtime::InteropServices::Marshal::FreeHGlobal(ptr);
    }
  }

Can anyone help here? Or any constructive criticisms you can offer?

Just in case you're curious: The project is for an ActiveRecordAdapter
using ADO.NET through C++/CLI, and is under 400 lines of C++; half of
it (Helpers.cpp) lifted from John Lam's RubyClr project. It seems to be
working relatively stably in testing, but the entire pont of the
project is to write something that properly handles the MSSQL nvarchar
(UTF-16LE) to Ruby UTF-8 String conversion (and vice-versa).