Using WideCharToMultiByte with Codepage 1258 (Vietnamese)

I am trying to convert a wide char string to Vietnamese ANSI (CP 1258), and for some reason half of the d开发者_高级运维iacritics are wrong. I did some research and unearthed this: http://www.siao2.com/2005/04/19/409566.aspx and this: http://www.siao2.com/2005/08/27/457224.aspx, which kind of confirm my suspicion that this is a special case. I added the WC_COMPOSITECHECK flag, but it does not help.

What I don't understand is why it is different from Thai, which also has this kind of multilayered diacritics. Thai worked just fine for me. Is there any kind of secret handshake I must use?