If I switch on my laptop BEFORE the mains-adaptor, when the system boots up and I login to a gnome desktop, the gnome-power-manager notification thingy in the panel will continually pop-up a "Battery Charged" (Your battery is now fully charged) message every 30 seconds or so. My laptop is a Dell D800, just over two years old. The battery is probably getting a bit ropey by now, but still gives me about 2.5 hours of uptime on it's own. The version of gnome-power-manager is gnome-power-manager-2.13.92-1 I have found that if I then kill the gnome-power-management process and start another one (actually, by selecting System -> Preferences -> More Preferences -> Power Management from the menus), the new process does NOT pop-up the "Battery Charged" notification (which is nice).
Can you get a verbose trace, and lshal -m output when the notification is flashing please? See http://www.gnome.org/projects/gnome-power-manager/bugs.html for how to get these.
Richard - in between raising this bug and your comment, I have performed a yum update. When I boot up without power, the power-manager icon (does a tooltip when I hover over it) states "Computer is running on battery power. Laptop battery 2 hours, 32 minutes remaining (96%)". **HOWEVER** when I plug the power cable in (in an effort to get these notifications going for you again), the gnome-power-manager does not realise this and thinks I am still running off battery !!! So I can't provide any tracing for you :-( This mornings updates (among other, I can give the full list if you want) were: gnome-power-manager.i386 2.13.92-1 -> 2.13.92-2 dbus.i386 FC5t3 -> 0.61-3 dbus-glib.i386 FC5t3 -> 0.61-3 dbus-x11.i386 FC5t3 -> 0.61-3 dbus-python.i386 FC5t3 -> 0.61-3 dbus-sharp.i386 FC5t3 -> 0.61-3 Here are the versions of the software I've currently got: gnome-power-manager-2.13.92-2 dbus-0.61-3 dbus-glib-0.61-3 dbus-python-0.61-3 dbus-sharp-0.61-3 dbus-x11-0.61-3 hal-0.5.7-1 hal-cups-utils-0.5.5-1.2 kernel-2.6.15-1.1977_FC5 kernel-devel-2.6.15-1.1977_FC5 Raw ACPI Info (after I plugged in the power connector) ====================================================== /proc/acpi/battery/BAT0/alarm alarm: 7159 mWh /proc/acpi/battery/BAT0/info present: yes design capacity: 71590 mWh last full capacity: 71590 mWh battery technology: rechargeable design voltage: 11100 mV design capacity warning: 7159 mWh design capacity low: 2169 mWh capacity granularity 1: 715 mWh capacity granularity 2: 715 mWh model number: DELL 0004P2 serial number: 10078 battery type: LION OEM info: Sony /proc/acpi/battery/BAT0/state present: yes capacity state: ok charging state: charging present rate: 32402 mW remaining capacity: 66650 mWh present voltage: 12762 mV /proc/acpi/battery/BAT1/alarm present: no /proc/acpi/battery/BAT1/info present: no /proc/acpi/battery/BAT1/state present: no So, I've killed off the hald and gnome-power-manager processes and started them in no-daemon verbose modes, both logging to a file. This is while the power cable is unplugged. Then I'll plug the power cable back in, leave things for a minute or so and then kill the processes again. I'll then attach the logs to this bug report.
Created attachment 125520 [details] Log file from HAL daemon
Created attachment 125521 [details] Log file from GNOME Power Manager
I see you have two batteries, both of which "sort of" fully charged. Can you reproduce this behaviour with only one battery installed?
Sorry Richard, but I've only got one battery (and always have had!). The Dell has room for two. I can withdraw the CD drive and replace it with a second battery pack. You can see that the second battery is "not present", above, by the lines: /proc/acpi/battery/BAT1/alarm present: no /proc/acpi/battery/BAT1/info present: no /proc/acpi/battery/BAT1/state present: no After my last comment, kept the laptop on all day and the Gnome Power Manager still insisted that the battery was discharging with 96% remaining :-)
2006-03-03 Richard Hughes <richard> * src/gpm-manager.c (battery_status_changed_primary): Remove the extra return that sneaked in to the notification path. This caused people to get repeated "battery is fully charged" messages if thier hardware was a little wonky. Should fix lp:32650 and rh:183612 Redhat also need to use the dbus 0.61 patch to g-p-m in rawhide (if it's going to work correctly), as the dbus glib api has changed. Yay. I'll push the g-p-m dbus-glib patch upstream now. Richard.
This is CVS. You *NEED* to use this patch if you want g-p-m to work on an freshly updated rawhide system. 2006-03-03 Richard Hughes <richard> * configure.in: Get the DBUS version, as now we have do stuff differently depending on whether we are using dbus 0.60 or dbus 0.61. Yay. I love API breakages. Sjoerd Simons found the bug, did the research, found the solution, and wrote a patch to fix, for which I thank him greatly: In dbus <= 0.60 dbus struct weren't really typed from dbus-glib's pov. In dbus >= 0.61 they are, so when defining a signal you need to be more precise then before. * src/gpm-hal-monitor.c (watch_device_connect_property_modified): get the correct struct type if we are using >= DBUS 0.61. Richard.
After today's yum updates: gnome-power-manager-2.13.93-1 hal-0.5.7-3 kernel-2.6.15-1.2009.4.2_FC5 kernel-devel-2.6.15-1.2009.4.2_FC5 everything seems to work fine. I can reboot without a power lead and g-p-m reports the remaining battery time - and this time-remaining changes over time. I can then re-plug-in the power lead and g-p-m updates to show the time-until-charged and this value also changes over time. Finally, when fully recharged, I get only one notification message. Great work Richard and Sjoerd!
Nice one!