On Fri, Jul 20, 2012 at 01:40:52 +0200, Vincent Lefevre wrote:

> On 2012-07-14 21:38:05 +0200, Vincent Lefevre wrote:
> > So, it seems that when the problem occurs, the keyboard modifiers may
> > still be working with clicks (to be confirmed).
>
> Forget that. The problem is the following: a keypress is taken into
> account only if the key is kept pressed for about half a second (key
> modifiers are generally pressed at least that time when used as a
> combination with a click, that's why they appear to be working as
> usual), a bit like the Power button needs to be pressed for some time
> to be taken into account. If a normal key is pressed long enough, it
> repeats until it is released, as usual.
>
Sounds like you have slowkeys enabled.
http://who-t.blogspot.fr/2012/06/xkb-slowkeys.html

Cheers,
Julien

On 2012-07-20 08:31:19 +0200, Julien Cristau wrote:
> Sounds like you have slowkeys enabled.
> http://who-t.blogspot.fr/2012/06/xkb-slowkeys.html

So, it would seem that some part of the system would enable SlowKeys
in my back for one of the keyboards (I recall that when this happens
while I'm using the USB keyboard, only the USB keyboard is affected,
not the laptop keyboard).

If there a way to know whether SlowKeys is enabled, for each available
keyboard? (Note: I'm not using GNOME, and even GNOME would be useless
because according to what I see on this page, it cannot differentiate
keyboards.)

--
Vincent Lefèvre <vincent@vinc17.net> - Web: <http://www.vinc17.net/>
100% accessible validated (X)HTML - Blog: <http://www.vinc17.net/blog/>
Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)


--
To UNSUBSCRIBE, email to debian-kernel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: 20120720065857.GA4794@xvii.vinc17.org">http://lists.debian.org/20120720065857.GA4794@xvii.vinc17.org

On 2012-07-20 08:58:57 +0200, Vincent Lefevre wrote:
> So, it would seem that some part of the system would enable SlowKeys
> in my back for one of the keyboards (I recall that when this happens
> while I'm using the USB keyboard, only the USB keyboard is affected,
> not the laptop keyboard).
>
> If there a way to know whether SlowKeys is enabled, for each available
> keyboard? (Note: I'm not using GNOME, and even GNOME would be useless
> because according to what I see on this page, it cannot differentiate
> keyboards.)

Additional information:

* SlowKeys can be turned on and off by pressing the Shift key for
at least 10 seconds:
https://bugzilla.redhat.com/show_bug.cgi?id=816764
I've tried and I confirm that this "works". When several keyboards
are attached, only the keyboard for which the Shift key is pressed
is affected (I suppose that this is expected).

* When the bug occurred in my case, I don't think I've pressed Shift
for 10 seconds (well, when I do this, this is in combination with
another key like PageUp / PageDown, but in this case, SlowKeys
switching isn't triggered), at least in most of the cases.

* It may happen that the system misses the Shift Release event.
Actually this is what happened on July 14. I recall:
Something interesting occurred with the USB keyboard. I was using
Iceweasel, and suddenly it behaved as if both the Shift and Ctrl
keys were pressed: left-clicks were extending the selection, and
a left-click on a link was opening it in a background tab. After
hitting various keys, this no longer occurred, [...]
And this is what happened last night. This could explain why the
problem occurs and that this could be specific to the USB keyboard
on this machine.

* I had said that the problem also occurred with the main keyboard,
but I don't remember whether I did anything special with the Shift
key in this case.

To summarize, there could be two different problems (but I'm not
sure):

1. The fact that pressing the Shift key for at least 10 seconds
activates SlowKeys for this keyboard. I don't think that this
feature should be enabled by default. Hidden features are bad.
Also, there may be a good reason to hold the Shift key for
10 seconds: when using it in a combination of Shift + clicks.
And when SlowKeys is enabled/disabled (in this way at least),
something should be written to some log file, so that the user
can get the information in case he wonders.

2. The fact that a Shift Release event can be missed (it remains
to be confirmed) with this USB keyboard.
Note: I haven't noticed any missing Release event for the other
keys.

--
Vincent Lefèvre <vincent@vinc17.net> - Web: <http://www.vinc17.net/>
100% accessible validated (X)HTML - Blog: <http://www.vinc17.net/blog/>
Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)


--
To UNSUBSCRIBE, email to debian-kernel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: 20120720074204.GB4794@xvii.vinc17.org">http://lists.debian.org/20120720074204.GB4794@xvii.vinc17.org

Vincent Lefevre <vincent@vinc17.net> writes:
> On 2012-07-20 08:58:57 +0200, Vincent Lefevre wrote:
>> So, it would seem that some part of the system would enable SlowKeys
>> in my back for one of the keyboards (I recall that when this happens
>> while I'm using the USB keyboard, only the USB keyboard is affected,
>> not the laptop keyboard).
>>
>> If there a way to know whether SlowKeys is enabled, for each available
>> keyboard? (Note: I'm not using GNOME, and even GNOME would be useless
>> because according to what I see on this page, it cannot differentiate
>> keyboards.)
>
> Additional information:
>
> * SlowKeys can be turned on and off by pressing the Shift key for
> at least 10 seconds:
> https://bugzilla.redhat.com/show_bug.cgi?id=816764
> I've tried and I confirm that this "works". When several keyboards
> are attached, only the keyboard for which the Shift key is pressed
> is affected (I suppose that this is expected).

Thanks a lot for posting this. I've now hit the bug several times
myself over the last few days, and without having read this thread first
I would never have known how to get out ot it. And that would have been
really annoying given that this typically happens in the middle of some
heavy editing session. I would have hated to have to throw away the work
and reboot to get the system working again, and I don't think I ever
would have figured out the shift key trick.

BTW, is it only me, or do you have to hold down the key significantly
longer to turn the "feature" off than to turn it on? It certainly feels
like it.

> * When the bug occurred in my case, I don't think I've pressed Shift
> for 10 seconds (well, when I do this, this is in combination with
> another key like PageUp / PageDown, but in this case, SlowKeys
> switching isn't triggered), at least in most of the cases.

In my case it seems to happen most often when I do cut and paste,
combining mouse selection with shift + del. I often end up holding down
the shift key while selecting and bam...


IMNSHO, the hotkey choice is too generic to be acceptable at
all. Holding down any single key to silently enable a feature like this
is not going to do. I find it somwhat surprising in light of the
Ctrl+Alt+Backspace disabling....

Why couldn't they have used *both* shift keys as a trigger? That would
have eliminated all false positives with no drawbacks AFAICS.

Anyway, please disable this extremely annoying feature until some safer
trigger can be found.


Bjørn


--
To UNSUBSCRIBE, email to debian-kernel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: 87r4ryjp3z.fsf@nemi.mork.no">http://lists.debian.org/87r4ryjp3z.fsf@nemi.mork.no

On 2012-07-26 21:31:44 +0200, Bjørn Mork wrote:
> BTW, is it only me, or do you have to hold down the key significantly
> longer to turn the "feature" off than to turn it on? It certainly feels
> like it.

No, it seems 10 seconds in both cases.

> > * When the bug occurred in my case, I don't think I've pressed Shift
> > for 10 seconds (well, when I do this, this is in combination with
> > another key like PageUp / PageDown, but in this case, SlowKeys
> > switching isn't triggered), at least in most of the cases.
>
> In my case it seems to happen most often when I do cut and paste,
> combining mouse selection with shift + del. I often end up holding down
> the shift key while selecting and bam...

Yes, while SlowKeys isn't turned on with Shift + non-modifier key,
if you hold Shift for 10 seconds while using the mouse, it will be
turned on.

> IMNSHO, the hotkey choice is too generic to be acceptable at
> all.

I completely agree.

--
Vincent Lefèvre <vincent@vinc17.net> - Web: <http://www.vinc17.net/>
100% accessible validated (X)HTML - Blog: <http://www.vinc17.net/blog/>
Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)


--
To UNSUBSCRIBE, email to debian-kernel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: 20120727001159.GF4993@xvii.vinc17.org">http://lists.debian.org/20120727001159.GF4993@xvii.vinc17.org

+++ Vincent Lefevre [2012-07-20 08:58 +0200]:
> On 2012-07-20 08:31:19 +0200, Julien Cristau wrote:
> > Sounds like you have slowkeys enabled.
> > http://who-t.blogspot.fr/2012/06/xkb-slowkeys.html
>
> So, it would seem that some part of the system would enable SlowKeys
> in my back for one of the keyboards

I hadn't seen this problem for weeks and thought maybe it was fixed,
and then it happened again (keyboard suddenly not responding, but
other keyboard working OK). Given the above message I tried holding
down shift for 10 seconds and lo and behold, the keyboard was back
working. So it does seem that the problem is the 'random but rare'
enabling of this feature without actually holding shift down for 10
seconds (or at least I'm fairly sure I didn't, but will keep an eye
out for doing it now).

A way to confirm if the slowkeys feature is currently engaged would
allow me to confirm this more directly next time.

Wookey
--
Principal hats: Linaro, Emdebian, Wookware, Balloonboard, ARM
http://wookware.org/


--
To UNSUBSCRIBE, email to debian-kernel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: 20120727013914.GE16745@stoneboat.aleph1.co.uk">htt p://lists.debian.org/20120727013914.GE16745@stoneboat.aleph1.co.uk

On Fri, Jul 27, 2012 at 02:39:14 +0100, Wookey wrote:

> A way to confirm if the slowkeys feature is currently engaged would
> allow me to confirm this more directly next time.
>
http://cgit.freedesktop.org/xorg/xserver/commit/?id=4af8e22b1a539778388fe509a7f3a25860a7879c
is in the X server in sid, so the X log should tell you.

Cheers,
Julien

Julien Cristau <jcristau@debian.org> writes:

> On Fri, Jul 27, 2012 at 02:39:14 +0100, Wookey wrote:
>
>> A way to confirm if the slowkeys feature is currently engaged would
>> allow me to confirm this more directly next time.
>>
> http://cgit.freedesktop.org/xorg/xserver/commit/?id=4af8e22b1a539778388fe509a7f3a25860a7879c
> is in the X server in sid, so the X log should tell you.

That's OK, but how do I permanently disable this mis-feature? I am now
starting to get pissed enough to be willing to replace whatever part of
the system is necessary. I can probably manage better without an X
server than without a keyboard...


Bjørn


--
To UNSUBSCRIBE, email to debian-kernel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: 87hass50fz.fsf@nemi.mork.no">http://lists.debian.org/87hass50fz.fsf@nemi.mork.no

On Sat, Jul 28, 2012 at 00:00:32 +0200, Bjørn Mork wrote:

> Julien Cristau <jcristau@debian.org> writes:
>
> > On Fri, Jul 27, 2012 at 02:39:14 +0100, Wookey wrote:
> >
> >> A way to confirm if the slowkeys feature is currently engaged would
> >> allow me to confirm this more directly next time.
> >>
> > http://cgit.freedesktop.org/xorg/xserver/commit/?id=4af8e22b1a539778388fe509a7f3a25860a7879c
> > is in the X server in sid, so the X log should tell you.
>
> That's OK, but how do I permanently disable this mis-feature? I am now
> starting to get pissed enough to be willing to replace whatever part of
> the system is necessary. I can probably manage better without an X
> server than without a keyboard...
>
As explained in the blog post I linked to, you have to disable AccessX.

Cheers,
Julien

On 2012-07-28 00:11:13 +0200, Julien Cristau wrote:
> As explained in the blog post I linked to, you have to disable AccessX.

But the blog post doesn't explain *how* to do that. Well, it explains
it only for GNOME 3 users. But not everyone uses GNOME.

"man -k accessx" gives nothing interesting.

http://tldp.org/HOWTO/Keyboard-and-Console-HOWTO-15.html says that
AccessX is enabled if the X server is run with the +accessx option,
but here AccessX is enabled while here there is no such option.

"grep -ri accessx /etc" and "ps -aef | grep -i accessx" give nothing.
And the X(7) man page doesn't mention AccessX.

I also wonder why such a feature is enabled by default while there
is no documentation about it!

--
Vincent Lefèvre <vincent@vinc17.net> - Web: <http://www.vinc17.net/>
100% accessible validated (X)HTML - Blog: <http://www.vinc17.net/blog/>
Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)


--
To UNSUBSCRIBE, email to debian-kernel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: 20120727230011.GA7143@xvii.vinc17.org">http://lists.debian.org/20120727230011.GA7143@xvii.vinc17.org