This is very difficult to reproduce manually, but it happens when running WebDriver tests where loads are very fast and multiple loads are done quickly. The problem is that we use an idle to resent the progress bar, but we don't reset it when a new load starts. We always reset the last idle on destroy, but if there's another one leaked, it will crash when scheduled if the window has already been destroyed. Thread 1 "MiniBrowser" received signal SIGSEGV, Segmentation fault. g_type_check_instance_cast (type_instance=0xaaaaaaaaaaaaaaaa, iface_type=94876889051072) at gtype.c:4052 4052 if (type_instance->g_class) (gdb) bt #0 g_type_check_instance_cast (type_instance=0xaaaaaaaaaaaaaaaa, iface_type=94876889051072) at gtype.c:4052 #1 0x0000564a3e32c345 in resetEntryProgress () #2 0x00007f2c58c79063 in g_timeout_dispatch ( source=source@entry=0x564a3f26ca00, callback=<optimized out>, user_data=<optimized out>) at gmain.c:4650 #3 0x00007f2c58c78615 in g_main_dispatch (context=0x564a3f006290) at gmain.c:3177 #4 g_main_context_dispatch (context=context@entry=0x564a3f006290) at gmain.c:3830 #5 0x00007f2c58c789b8 in g_main_context_iterate (context=0x564a3f006290, block=block@entry=1, dispatch=dispatch@entry=1, self=<optimized out>) at gmain.c:3903 #6 0x00007f2c58c78cc2 in g_main_loop_run (loop=0x564a3f52fbf0) at gmain.c:4099 #7 0x00007f2c5a834e25 in gtk_main () at gtkmain.c:1323 #8 0x0000564a3e32567c in main ()
Created attachment 340687 [details] Patch
Comment on attachment 340687 [details] Patch Attachment 340687 [details] did not pass mac-wk2-ews (mac-wk2): Output: http://webkit-queues.webkit.org/results/7721376 New failing tests: transitions/created-while-suspended.html
Created attachment 340689 [details] Archive of layout-test-results from ews104 for mac-sierra-wk2 The attached test failures were seen while running run-webkit-tests on the mac-wk2-ews. Bot: ews104 Port: mac-sierra-wk2 Platform: Mac OS X 10.12.6
Committed r232009: <https://trac.webkit.org/changeset/232009>
<rdar://problem/40407917>