> I see what the problem is now. packet_list_select_cb() is using the row > passed to it by GTK+ as the index into the array of frame_data structs. > But with my patch, when filtered, the row does is not the same as the > packet number. > > No fix yet, but at least I know what the problem is. When doing the changes to allow the time format to be changed on the fly, I'd thought about this one; one possibility would, I think, be to store with each entry in the clist for the packet summaries a pointer to the "frame_data" structure for the frame in question and, in "packet_list_select_cb()", do: fd = (frame_data *)gtk_clist_get_row_data(GTK_CLIST(w), row); to get the appropriate "fd" value.
Powered by MHonArc 2.6.10