From bruen...@google.com on February 21, 2013 23:13:43

running mingw gtk3-demo (xref issue #1046 ) there are some uninits that are
probably false positives, and a lot of possible leaks:

Dr.M ERRORS FOUND:
Dr.M 1 unique, 1 total unaddressable access(es)
Dr.M 40 unique, 505 total uninitialized access(es)
Dr.M 0 unique, 0 total invalid heap argument(s)
Dr.M 4 unique, 38 total GDI usage error(s)
Dr.M 0 unique, 0 total warning(s)
Dr.M 21 unique, 62 total, 7693 byte(s) of leak(s)
Dr.M 355 unique, 379 total, 16450 byte(s) of possible leak(s)
Dr.M ERRORS IGNORED:
Dr.M 12796 still-reachable allocation(s)
Dr.M (re-run with "-show_reachable" for details)

most are like this:
Error #414: POSSIBLE LEAK 36 direct bytes 0x01ff0e70-0x01ff0e94 + 0 indirect bytes
#0 libglib-2.0-0.dll!g_malloc0 [../../glib/gmem.c:189]
#1 libgobject-2.0-0.dll!g_closure_new_simple [../../gobject/gclosure.c:206]
#2 libgobject-2.0-0.dll!g_cclosure_new [../../gobject/gclosure.c:917]
#3 libgobject-2.0-0.dll!g_signal_new_class_handler [../../gobject/gsignal.c:1480]
#4 libgtk-3-0.dll!gtk_text_view_class_intern_init [../../gtk/gtktextview.c:929]
#5 libgobject-2.0-0.dll!g_type_class_ref [../../gobject/gtype.c:2239]
#6 libgobject-2.0-0.dll!g_object_newv [../../gobject/gobject.c:1624]
#7 __tmainCRTStartup [../crt/crtexe.c:315]
#8 KERNEL32.dll!BaseThreadInitThunk
#9 ntdll.dll!__RtlUserThreadStart

I suspect glib's extra layer is involved here.

logs are attached.

logs are in ~/drmemory/bugs/i1046-gtk/DrMemory-gtk3-demo.exe.7948.000/
and /work/drmemory/bugs/i1046-gtk/DrMemory-gtk3-demo.exe.7948.000/

Original issue: http://code.google.com/p/drmemory/issues/detail?id=1143