> >Hmmm, if I'm right Unicode can have null bytes in some chars, so Unicode >strings should be terminated by a double '\0', shouldn't they ? > >my 0.02 Euros before going to bed, >Fran=E7ois. well... depends on what you mean by "double". Unicode strings are composed of wide characters (type wchar_t) which are two bytes under most implementations (but this is not a requirement -- they could be 4 bytes or longer). So the ending null is still just "one" character (however many bytes that is). For example sizeof(L"abc") =3D=3D 8. That's four characters, including the terminating null, each of which is two bytes wide (in BeOS Intel, wchar_t is a typedef for unsigned short).