Comment 8 for bug 745115

These traces are uniform but bizarre. They all seem to be during the processing of an ApplicationIconChanged signal, but the stack trace shows the local variables as if it were handling a ApplicationLabelChanged signal, and even shows a application_label_changed() call (with the wrong function signature) in the stack.

But the variable contents and the signal name itself point to ApplicationIconChanged being the actual signal. I think I trust those data points slightly more than I trust gdb's after-the-fact processing of the stack. That combined with the bottom frame always missing points to some stack corruption. It's just odd that it's so uniform across the duplicates.