For the first time I can recall, Alexandre committed lots of changes on a Saturday, so I did a valgrind run.
http://kegel.com/wine/valgrind/logs-2007-12-08-summary.txt shows the highlights; basically, lots more warnings like + Syscall param writev(vector[1]) points to uninitialised byte(s) For example, the one at the top of http://kegel.com/wine/valgrind/logs-2007-12-08/vg-user32_edit-diff.txt
+ Syscall param writev(vector[1]) points to uninitialised byte(s) + at (within /lib/ld-2.5.so) + by writev (in /lib/tls/i686/cmov/libc-2.5.so) + by wine_server_call (server.c:244)
The only change in the server was http://source.winehq.org/git/wine.git/?a=commitdiff;h=edd3f264d95c0c6bed9c18... so I guess the new errors reflect additional error completion events being sent, and that kind of event just happens to contain some uninitialized bytes.
So this is probably in Alexandre's court, and my guess is he'll consider this a false positive. - Dan