31 Oct. Found that bluetooth keyboards become even more prone to repeats when
using a Nokia LD-1W GPS receiver. The GPS works fine with Maemo Mapper, but the
keyboards (Apple Wireless, Stowaway and Nokia SU-8) become even more difficult
to use.

Why is there such an interaction? How can it be remedied in the current OS?

*** This bug has been confirmed by popular vote. ***

confirmed also using stowaway bt keyb. while connected with a bt phone...
FIX this thanks !

Comment 4 Roope Rainisto 2007-11-05 10:00:42 UTC

Adding CC.

from what I have seen in 2008 so far, this still seems like a problem. Any
hints as to where one should look to try to fix this?

Comment 6 Tomas Junnonen 2007-11-16 13:12:34 UTC

In IT2008 you can modify the repeat rate for the BT keyboard through gconf:

"gconftool-2 -s /apps/osso/inputmethod/ext_kb_repeat_delay -t int 800"
This sets the delay (in msec) before key repeat starts

"gconftool-2 -s /apps/osso/inputmethod/ext_kb_repeat_interval -t int 100"
This sets the time interval (in msec) between keys when repeating

However, I'm not so sure this problem is related to the repeat rate itself,
although you can check if bumping the values (for example to 1000/150, or even
much more) helps. Probably the presence of other BT devices is the triggering
factor.

sorry, the Connection status bar applet is for DUN, not generic Bluetooth.

Neither setting the gconf value nor disabling X's autorepeat (i.e.
XAutoRepeatOff()) solves this problem. But setting the repeat rate directly on
the event input device does work. evrepeat in initfs appears to work. Running
"chroot /mnt/initfs evrepeat /dev/input/eventN 5000 100" sets the delay to 5s
and eliminates problems with flaky bluetooth connections leading to unwanted
repeats. I made this permanent by adding to /etc/udev/rules.d the file
010_keyboard-repeat.rules with content:
KERNEL=="event[0-9]*", SUBSYSTEMS=="bluetooth", RUN+="/usr/sbin/chroot
/mnt/initfs evrepeat /dev/input/%k 5000 100"

> KERNEL=="event[0-9]*", SUBSYSTEMS=="bluetooth", RUN+="/usr/sbin/chroot
> /mnt/initfs evrepeat /dev/input/%k 5000 100"

mmm no, it isn't solved , at least on my n800....
still random repeats occurs even if lower than before....

Austin, thanks! Works well for me (tested on a N800 running 4.2007.38-2)

aaargh :)
i've just realized that i've missed the final '"' :)
I'm testing but seems to work well for now :)
Thanks for the great solution !!

Shouldn't this remain open until it is actually fixed? Seems like evrepeat
should be fixed to a more appropriate value, or there should be a way for a
normal user to adjust it.

Thanks to Bob for discovering that the previous method doesn't continually work
due to the lack of dev files under /mnt/initfs. Once a keyboard has been cycled
a few times, there won't be a correct dev file at /mnt/initfs/dev/input/eventN.
There are several solutions to this. One: compile evrepeat or something similar
outside of initfs (would require the source or some hacking). Two: add "mount
--bind /dev /mnt/initfs/dev" somewhere to your startup scripts. Three: modify
the udev run command to copy over the dev file before doing the chroot, e.g. cp
/dev/input/%k /mnt/initfs/dev/input. I've only tested #2.

Also see enhancement request at bug 2670
https://bugs.maemo.org/show_bug.cgi?id=2670

I was attempting to use a Stowaway BT keyboard on a crowded train this weekend
(N800, OS2008 2.2007.50-2), and this bug was neither resolved nor "works for
me". Increasing the repeat rate works around the problem.

I threw together a package to install a workaround automagically (provides an
evrepeat replacement), get it at http://apt.notcom.org/install/btrepeat.install
or http://apt.notcom.org/pool/user/b/btrepeat/ for deb/source. Please test and
report if it works/doesn't work for you.

(In reply to comment #14)
> Please test and report if it works/doesn't work for you.

It works for me (latest 51_3 build of os2008 on n800).
Any chance you could add the second parameter, for repeat rate?  ^^

Nice to hear. Unfortunately, trying to change the repeat rate setting appears
to have no effect, so it's not included in the utility.

Comment 17 Frantisek Dufka 2008-07-28 10:36:13 UTC

(In reply to comment #13)
> . Three: modify
> the udev run command to copy over the dev file before doing the chroot, e.g. cp
> /dev/input/%k /mnt/initfs/dev/input.

Until Diablo, initfs was not writable, but one can use /tmp which is shared
between rootfs and initfs.

BTW, here is code snippet from http://fanoush.wz.cz/maemo/#kbdd that changes
repeat rate for keyboard named "SerKBD":

#change key repeat rate
cd /sys/class/input
for dev in input* ; do
        if [ "$(cat $dev/name)" = "SerKBD" ] ; then
                cd $dev
                for ev in event* ; do
                        cp -a /dev/input/$ev /tmp/$$-$ev
                        #this works because /tmp is shared with initfs
                        chroot /mnt/initfs evrepeat /tmp/$$-$ev
$KEYREPEAT_DELAY $KEYREPEAT_PERIOD
                        rm /tmp/$$-$ev
                done
        fi
done

How long has this issue been open? Previous solutions don't seem to work and if
it wasn't that I've come to rely on the gps for work I do, I would have dumped
this by now as it is next to useless for extended typing unless the repeat rate
is reduced. 

I am using N810 with Diablo 4.2008.36-5. Apple BT keyboard (and have also tried
it with a Stowaway).


Wwwwwwe need a definite solution from Nokia on this. AAll too long they have
been avoiding addressing this.

Comment 19 Andre Klapper 2008-10-13 12:47:21 UTC

At least with the Outside/iGo (Stowaway Ultra-Slim) BT keyboard there's a lot
of issues out there, see e.g. bug 3532.

N810 with 5.2008.43-7 and SU-8W the instant key repeat problem is still there.
Ubuntu 8.10 has a similar problem with it's USB keyboard that the set kbdrate
is changed during the boot process even when specificaly set by kbdrate. Once X
is started I have to run kbdrate again. I don't know if this is only a
coincidence.

Comment 21 Andre Klapper 2009-02-03 01:30:40 UTC

Updating version info; Setting priority to low (sorry, looks like the
developers currently have other priorities)

eero: could you possibly test this w/ a fremantle stack? the internal bug is
currently marked against diablo....

Comment 23 Eero Tamminen 2009-06-03 10:14:43 UTC

(In reply to comment #22)
> eero: could you possibly test this w/ a fremantle stack? the internal bug is
> currently marked against diablo....

Don't have a BT keyboard & not area & no time.

(In reply to comment #14)
> I was attempting to use a Stowaway BT keyboard on a crowded train this weekend
> (N800, OS2008 2.2007.50-2), and this bug was neither resolved nor "works for
> me". Increasing the repeat rate works around the problem.
> 
> I threw together a package to install a workaround automagically (provides an
> evrepeat replacement), get it at http://apt.notcom.org/install/btrepeat.install
> or http://apt.notcom.org/pool/user/b/btrepeat/ for deb/source. Please test and
> report if it works/doesn't work for you.
> 

Thank you very much- I installed the deb-Package an it works on my n810 with
OS2008 and a hp-foldable-bt-keyboard. You solved a problem that was driving me
crazy for months now!

Comment 25 Andre Klapper 2009-09-28 16:40:29 UTC

This is a WONTFIX for both Maemo5 and Maemo4 as using a Bluetooth keyboard is
not officially supported in Maemo 5 (of course improvements can always be made
by community or third party enhancements).

Please note though that according to
http://wiki.maemo.org/Fremantle_Unsupported_Bluetooth_profiles Maemo5 comes
with the bluetoothd input plugin installed but disabled through
/etc/bluetooth/main.conf.