wcslen() returns the number of characters in a wchar_t string.
In the case of WideCharToMultiByte() an estimate for the size of
the utf8 buffer is needed. Using wcslen() is incorrect for such a buffer,
because for any non-ASCII character the estimate will be off by 1 byte.
Call the following instead to obtain the proper UTF8 buffer size
for the conversation:
WideCharToMultiByte(CP_UTF8, 0, utf16, -1, NULL, 0, NULL, NULL);
Also fix some missing "\n" in fprintf() calls.
Signed-off-by: Lubomir I. Ivanov <neolit123@gmail.com>