"L. Rahyen" research@science.su wrote:
This problem was already solved in this patch [1] by Nigel Liang. Unfortunately, it wasn't accepted and Nigel didn't tried to ask why it was rejected or to resend it. I think I will go ahead and resend his patch.
[1] http://www.winehq.org/pipermail/wine-patches/2007-October/044977.html
MultiByteToWideChar is that patch passes negative length for the destination buffer which is incorrect.