GNOME Bugzilla – Bug 428007
leak in spi_atk_bridge_focus_tracker
Last modified: 2007-04-12 07:10:54 UTC
valgrind tells me: ==3266== 133 (108 direct, 25 indirect) bytes in 3 blocks are definitely lost in loss record 77 of 196 ==3266== at 0x4004824: calloc (vg_replace_malloc.c:279) ==3266== by 0x6AB9BD: g_malloc0 (gmem.c:150) ==3266== by 0x5BE4A65: ORBit_alloc_by_tc (in /usr/lib/libORBit-2.so.0.1.0) ==3266== by 0x5BDF9AC: ORBit_small_alloc (in /usr/lib/libORBit-2.so.0.1.0) ==3266== by 0x45BF067: spi_init_any_nil (util.c:124) ==3266== by 0x400BAC6: spi_atk_bridge_init_nil (bridge.c:1340) ==3266== by 0x400C1BF: spi_atk_bridge_focus_tracker (bridge.c:706) ==3266== by 0xB92BCB: atk_focus_tracker_notify (atkutil.c:199) ==3266== by 0x452C5EF: gail_focus_idle_handler (gail.c:578) ==3266== by 0x6A28C0: g_idle_dispatch (gmain.c:3928) ==3266== by 0x6A4621: g_main_context_dispatch (gmain.c:2045) ==3266== by 0x6A75FE: g_main_context_iterate (gmain.c:2677) ==3266== I assume this is the dupe'd name field in the any. I never quite understood how the memory management for stack-allocated Accessibility_Event structs is supposed to work in bridge.c...
Matthias, which version of at-spi are you using?
1.18.1
Created attachment 86222 [details] [review] for fixing bug 428007 spi_atk_bridge_focus_tracker() does not free Accessibility_Event's any_data. Try this patch, thx.