Hi,
Profiling wireshark shows that new_packet_list_thaw()/new_packet_list_freeze()
takes ~64% of CPU time while loading file.
Removing some lines from file.c/cf_read:
@@ -607,10 +607,6 @@ cf_read(capture_file *cf, gboolean from_save)
if (progbar_quantum > 500000 || displayed_once == 0) {
if ((auto_scroll_live || displayed_once == 0 || cf->displayed_count < 1000) && cf->plist_end != NULL) {
displayed_once = 1;
- new_packet_list_thaw();
- if (auto_scroll_live)
- new_packet_list_moveto_end();
- new_packet_list_freeze();
}
}
speeds up loading capture file with 122 K packets from 10.908s to 4.118s
The loading time might be problem with my system or GTK+ (version 2.20.1),
and it'd be great if someone find time to test it.
About actual bug:
In old packet list _freeze stops GtkClist from redraw while adding packets to list.
In new packet list _freeze sets empty (NULL) model, and _thaw restores model with packets.
I really can't remember how old packet list works while loading capture file, but I think
code from cf_read was intented to show some packets (by redrawing GtkClist).
With new packet list no packets are shown, cause _freeze works completly different.
Ok to remove/comment out this code?
Cheers.
