valgrind regression in gdiplus

[Dan Kegel]

Yesterday, there were a bunch of new valgrind warnings in gdiplus,
and they're still there today.  See
http://kegel.com/wine/valgrind/logs-2008-06-26/vg-gdiplus_font.txt
For instance,
 Invalid read of size 4
    at  GdipCreateFont (font.c:99)
    by  test_createfont (font.c:57)
    by  func_font (font.c:250)
    by  run_test (test.h:449)
    by  main (test.h:498)
  Address 0x4 is not stack'd, malloc'd or (recently) free'd
...
 Conditional jump or move depends on uninitialised value(s)
    at  winetest_vok (test.h:272)
    by  winetest_ok (test.h:301)
    by  test_createfont (font.c:61)
    by  func_font (font.c:250)
    by  run_test (test.h:449)
    by  main (test.h:498)
  Uninitialised value was created by a stack allocation
    at  test_createfont (font.c:43)
...
 Syscall param write(buf) points to uninitialised byte(s)
    at  (within /lib/ld-2.7.so)
    by  (within /lib/tls/i686/cmov/libc-2.7.so)
    by  _IO_file_xsputn (in /lib/tls/i686/cmov/libc-2.7.so)
    by  (within /lib/tls/i686/cmov/libc-2.7.so)
    by  vfprintf (in /lib/tls/i686/cmov/libc-2.7.so)
    by  winetest_vok (test.h:279)
    by  winetest_ok (test.h:301)
    by  test_createfont (font.c:65)
    by  func_font (font.c:250)
    by  run_test (test.h:449)
    by  main (test.h:498)
  Address 0x7f21d657 is on thread 1's stack
  Uninitialised value was created by a stack allocation
    at  test_createfont (font.c:43)