Notices

Originally Posted by Mentalist Traceur

That's unfortuante. Do you mean literally sunk? Утонуло? Or are you using that to mean broke/died/stopped working?

Originally Posted by Mentalist Traceur

Ironically enough I was going to report a bug regarding this applet - when using the Power Kernel, combined with the method for permanently enabling that Kernel's own battery reporting module ("echo bq27x00_battery >> /etc/modules" as root), I find that Advanced Power under-reports the actual battery percentage left value (More on the Power Kernel's battery reporting here: http://wiki.maemo.org/Kernel_Power#Battery ). By a rather inconsistent amount - 5-20% difference, at various times. I am not sure if this means it was always under-reporting, or if it is only because I have the bq27x00_battery module loaded that the problem occurs. For what it's worth, other battery reporting programs agree with the Advanced Power percentage. HealthCheck, and using the line "${execi 30 hal-device bme |grep "percentage" |awk '{print $3;}'}" in Conky. NetMon has a battery stats page in its program, but Netmon simply fails to load anything on the battery, showing ?s for every value ever since I made bq27x00 load at boot.

Now, here you might be thinking it's the kernel module that's wrong - and it probably is linked to stuff going haywire - but I have repeatedly (I noticed this around a week ago by now, been waiting to see if anyone else reported this or if my thing did nothing) been able to use the N900 for somewhere around 30 minutes to an hour (closer to two hours when on-off using it or not connected to wifi) with the battery icon being completely red, Advanced Power reporting 0.4 percent battery life, and HealthCheck and the above Conky command showing 0% battery life. Furthermore, after a full charge, at best, each of them displays a percentage in the mid 90s (sometimes as low as 84-86), while the Kernel Power's bq27x00 report is 100% after a full charge. Depending on use, the values reported go between 5 (right after a long full charge) to 20% or more of a gap. (With a trend towards the gap increasing as the battery discharges.)

Ultimately, this means my batter icon goes red on me when I still have at least an hour very-active-use of battery life left, or approx. 20% capacity.

Originally Posted by Mentalist Traceur

Since you no longer have an N900, I understand you can't really develop or test these applets. Any insight/help you can provide would be great, though.

Originally Posted by Mentalist Traceur

In the meantime, these are open source, correct? I'd be happy to try to dive in to the world of computer programming by trying to pick this and the Advanced Interface Switcher up and continuing from whatever your soon-to-be released versions are. Means progress will halt for a while while I try to make sense of everything, since I have almost no coding language familiarity at all right now, but at least these great applets don't have to stop improving.

Originally Posted by DaSilva

It sounds like a problem which should be easy to solve. Have you tried to flash your N900 with the current firmware? If there is no problem with the hardware it should work after flashing.

It's not the AP (in fact APM) bug. It's the difference in Nokia's BME and bq27x00 module. To make APM support power-kernel's battery module new wrapper should be written. It's quite easy

Well, I can develop a bit ) Just because a lot of things are already made, implementing new features is not that hard. But no opportunity to test (

In fact, it's quite easy, 'cause all the stuff is Python. Feel free to ask I still can develop a bit and help to improve

Originally Posted by 412b

What's the output of apmd.log?
And also try just starting job from xterm as root:

Code:

start apmonitord

Originally Posted by Mentalist Traceur

Hmmm... I'll see if rmmod bq27x00_battery makes the proper value get reported. Of course, with only one module reporting battery life the only way I can really test how accurate it is is time how long it takes for it to die after it reaches "0%".

Do you know anything about why the two kernel modules (if BME is a kernel module by definition - not sure) conflict? Does BME just always interact with the power kernel poorly, or does the bq27x00 module running at the same time as BME conflict?

Originally Posted by Mentalist Traceur

Is there any documentation or technical details on why the two modules cause problems that you either know or can direct me to? (In the meantime, I'm reading through this entire thread. I'm at page 15 right now, so it'll take a while before I get through the entire thing for any useful info it may have.)

Originally Posted by Mentalist Traceur

As for writing a new wrapper, I myself don't really know where to start with that, but I suspect that by the time I get around to being familiar enough with python for that to make a difference, I will know enough to figure it out.

Originally Posted by Mentalist Traceur

Well, I'm happy to keep testing stuff, and to give you as detailed of reports on the results as possible, if you wish to continue.

I can just get the current source from the garage page or through apt-get, right? Do you still plan on releasing the new features you were working on before your N900 died?