New valgrind warnings in kernel32/tests/codepage.c

[Colin Finck]

Dan Kegel wrote:
> your change [...] seems to have triggered some valgrind warnings,

Thanks for providing the valgrind output!

Indeed, I was passing a wrong length to a WideCharToMultiByte call, which 
led to the use of uninitialized stack memory.
Fixed here: 
http://www.winehq.org/pipermail/wine-patches/2008-July/058951.html

Best regards,

Colin