> (I'm not sure how I managed to reduce "g_list_append()" CPU time to > almost nothing, given that the code to add a row to a GtkClist uses it; > I'm obviously missing something. Yes, I'm missing the fact that, apparently, the folks who implemented the GtkClist also realized that passing "g_list_append()" a pointer to the first element of a GList, when constructing a list by appending stuff to it, isn't a very good idea, and keep around a pointer to the end of the GList and append after that instead....
Powered by MHonArc 2.6.10