I've edited the patch to use g_set_error instead of g_set_error_literal (simply replacing the function since the arguments seems to already be in the right format).
Since I have minimal knowledge about the actual code involved in this I can only confirm that it works with no immediate regressions on an i386 virtualbox instance.
Yet another debdiff attached, and testing packages available in my ppa.
I've edited the patch to use g_set_error instead of g_set_error_literal (simply replacing the function since the arguments seems to already be in the right format).
Since I have minimal knowledge about the actual code involved in this I can only confirm that it works with no immediate regressions on an i386 virtualbox instance.
Yet another debdiff attached, and testing packages available in my ppa.