Today's valgrind run shows a bunch more valgrind warnings in gdi32/tests/font.c: http://kegel.com/wine/valgrind/logs-2008-04-04/vg-gdi32_font-diff.txt e.g.
+ Use of uninitialised value of size 4 + at HEAP_CreateFreeBlock (heap.c:502) + by HEAP_MakeInUseBlockFree (heap.c:538) + by RtlFreeHeap (heap.c:1294) + by GDI_FreeObject (gdiobj.c:731) + by FONT_DeleteObject (font.c:554) + by DeleteObject (gdiobj.c:854) + by enum_truetype_font_proc (font.c:1602) + by FONT_EnumInstance (font.c:589) + by WineEngEnumFonts (freetype.c:3761) + by FONT_EnumFontFamiliesEx (font.c:626) + by EnumFontFamiliesExA (font.c:664) + by func_font (font.c:1744) + by run_test (test.h:406) + by main (test.h:455) ... + Conditional jump or move depends on uninitialised value(s) + at RtlAllocateHeap (heap.c:1204) + by WineEngGetOutlineTextMetrics (freetype.c:4831) + by WineEngGetTextMetrics (freetype.c:4720) + by GetTextMetricsW (font.c:1158) + by X11DRV_XRender_SelectFont (xrender.c:433) + by X11DRV_SelectFont (xfont.c:3216) + by FONT_SelectObject (font.c:497) + by SelectObject (gdiobj.c:1129) + by enum_truetype_font_proc (font.c:1579) + by FONT_EnumInstance (font.c:589) + by WineEngEnumFonts (freetype.c:3761) + by FONT_EnumFontFamiliesEx (font.c:626) + by EnumFontFamiliesExA (font.c:664) + by func_font (font.c:1744) + by run_test (test.h:406) + by main (test.h:455)
I don't think these are actually new; errors like them have been around, they just got triggered more with a new test that was added. It always worries me when heap routines start showing up in valgrind dumps, though; it'd be nice if somebody could take a look at this. - Dan