Comment 8 for bug 33523

Revision history for this message
Paul Sladen (sladen) wrote :

This probably explains where the random fade-outs are coming from; lots of extra timers that aren't getting destoryed:

  $ grep -c 'killing idle_timer' gnome_screensaver_debug_xLIHlJ
  536
  $ grep -c 'starting idle_timer' gnome_screensaver_debug_xLIHlJ
  582

eg. Here's a snippet during/around a suspend:

  [add_idle_timer] gs-watcher-x11.c:474 (02:02:24): starting idle_timer (50000, 454)
  [_gs_watcher_check_pointer_position] gs-watcher-x11.c:1389 (02:02:24): Idle 0 seconds
  ...
  [_gs_watcher_check_pointer_position] gs-watcher-x11.c:1389 (02:03:08): Idle 44 seconds
  [check_for_clock_skew] gs-watcher-x11.c:1308 (02:03:08): checking wall clock for hibernation, changed: 0:00:02
  XXX this is gap of 13 seconds whilst suspended
  [_gs_watcher_check_pointer_position] gs-watcher-x11.c:1389 (02:03:21): Idle 57 seconds
  [check_for_clock_skew] gs-watcher-x11.c:1308 (02:03:21): checking wall clock for hibernation, changed: 0:00:13
  [idle_timer] gs-watcher-x11.c:1239 (02:03:21): in idle timer
  [_gs_watcher_check_pointer_position] gs-watcher-x11.c:1389 (02:03:21): Idle 57 seconds
  [check_for_clock_skew] gs-watcher-x11.c:1308 (02:03:21): checking wall clock for hibernation, changed: 0:00:00
  [add_idle_timer] gs-watcher-x11.c:474 (02:03:21): starting idle_timer (3000, 455)
  [watcher_idle_notice_cb] gs-monitor.c:126 (02:03:21): Idle notice signal detected: 1

The ''wall clock has jumped by ..." code never seems to notice...