New valgrind regressions in riched20, rpcrt4

[Dan Kegel]

http://kegel.com/wine/valgrind/logs-2007-12-17-summary.txt
shows interesting things happened since Friday in riched20.
See
http://kegel.com/wine/valgrind/logs-2007-12-17/vg-riched20_editor-diff.txt
for a whole heap of new errors.  The only change since Friday
was Maarten's, http://www.winehq.org/pipermail/wine-cvs/2007-December/038829.html,
which seems somewhat unrelated, maybe he's getting slagged
unfairly here.  (But it would still be nice if he could have a look.)

There's also the somewhat questionable new warning
+ Source and destination overlap in memcpy(0x7F00F844, 0x7F00F844, 22)
in
http://kegel.com/wine/valgrind/logs-2007-12-17/vg-rpcrt4_ndr_marshall-diff.txt
which would be Rob's if it turns out to be a real problem.
- Dan