you can try any other system ?

Before flashin why not try to reinstall it either thru ham, fap, or try "apt-get install --reinstall" thru xterm

that's a good idea .

I'd wish so, but actually i can't - lan house is strongly limited and friends ith computers just too far away.
I flashed by Windows XP once... Actually the install cd is broken and i bought a cheap 2003. Headache included ;-)

Actually i prefer flashing cause i have a lot of rubbish over there. Contacts are saved anyway...

Ah i tried, but apt-get seems to be broken anyway on my phone. Reflash it. However. Better than struggeling with that **** more time around ;-)

Not exactly. This way, You don't have chance to properly debug the problem, then avoid it next time OR help improving community made apps / fix nokia bugs. Debugging need much patience, but fixing something Yourself, or as a cooperative work with other community members, is very rewarding. Trust me.

Anyway, of course, it's always Your decision what to do with Your device.

---

Anyway, You mentioned that You charge Your battery externally. The bq27200 module/chip can't properly detect events such as "new battery", so problem would probably disappear after charging device via USB. Which, obviously, You wasn't able to try, cause broken USB port. BME was reporting correct values, cause it calculates capacity somehow (including guesswork) via battery voltage. BME does NOT use raw bq chip measuring means.

I'm 99% sure, that somehow, colin's command to blacklist b1 module wasn't working for You. Which is no surprise for me, cause blacklisting doesn't work for me also Attention! I don't mean, that it's colin's fault - his command is perfectly fine. I tried every single method to blacklist bq module, checked n times, that my blacklist list is out of typos - to no avail. Only one way for me was to modprobe -r, which result in device restart ~32 minutes after executing.

Doing proper bq learning cycle (letting battery voltage as reported by bq to get under 3248 mV, without device turning off via bme, then charging to full), bq module work fine for me - at least, meaning that it doesn't cause problems (values returned are still somehow wrong, but i think I'm close to developing "user behavior hack" to fix that), so i just let it go.

Then, pyBattery (which is excellent piece of code, all-together with advanced-power-monitor by the way), always use bq data, when it's present. That's why Your applet always thought that battery is 0%. No big deal, mind You - bq module perform read only, so this was only visual mismatch, no risk of loosing anything, and certainly, didn't deserve reflashing device.

Hope it helps, at least for other people that may come here, searching for solution

Actually even when i charged by USB (yes, it's fixed) does not help :-(

Finally funny was what it worked a lot of times ago (yes, even by seperate recharge), i use the nokia dt-33 for that. I guess it could be some update of libraries
so that the error could appear. However, finally i killed apt-get by deleting too much files so i'm not even able to reinstall it by dpkg ...

Can You please execute:

cat /sys/class/power_supply/bq27200-0/capacity

And post Your output here?

There might be a workaround solution with broken microUSB port:
http://talk.maemo.org/showthread.php?t=75318

If you're handy with soldering tools and have the necessary equipment, you might be able to get alternative USB port to work and maybe even better, flash mode.