Stuck in always "clicking" state with QT5 apps on a Wacom Cintiq 27QHD Touch

classic Classic list List threaded Threaded
6 messages Options
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Stuck in always "clicking" state with QT5 apps on a Wacom Cintiq 27QHD Touch

Camille Bissuel
Hi dear Wacom users list,

In a desperate attempt, I'm reporting a nasty strange bug to your reading…

I suppose you all now Krita 3 is released, and that's a nice release ;)
Krita 3 is based on QT 5, and is supposed to have multi-touch support.
I encounter a very nasty bug with Krita 3 : I get stuck in an always clicking state with QT5 apps, meaning when the bug occur, I'm clicking everywhere by simply moving the cursor hover any interface. So everything become unusable and I have to log off my session or reboot to temporarily resolve it.

I already filled a lengthy bug report on Krita bug tracker : https://bugs.kde.org/show_bug.cgi?id=363225
But as it's clearly an upstream bug now, and as Krita team is already overwhelmed, I'm looking for some guidance on how to solve that and being able to use my new shiny tablet with Krita again ;)

I know there are very clever people on this list, and some that are particularly aware of the Linux input stack… I don't really know if the bug is coming from X server, from Wacom drivers or from QT5… and I'm not a developer… please help if you can !

My best wishes, and many thanks for your time and attention in advance.
Camille Bissuel - nylnook.com

------------------------------------------------------------------------------
What NetFlow Analyzer can do for you? Monitors network bandwidth and traffic
patterns at an interface-level. Reveals which users, apps, and protocols are
consuming the most bandwidth. Provides multi-vendor support for NetFlow,
J-Flow, sFlow and other flows. Make informed decisions using capacity
planning reports. https://ad.doubleclick.net/ddm/clk/305295220;132659582;e
_______________________________________________
Linuxwacom-discuss mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/linuxwacom-discuss
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Stuck in always "clicking" state with QT5 apps on a Wacom Cintiq 27QHD Touch

Jason Gerecke
Camille,

It would be good to know if this issue is introduced in X or Qt so that
we can know what layer is broken. I believe the following command should
be able to straightforwardly show the press/release events occurring at
the X layer:

$ xinput test-xi2 --root | awk -v OFS='\t' '/Press|Release/ {X=$0;
getline; print strftime("%H:%M:%S"), X, $0; fflush()}' | tee button.log

Go ahead and run that in a terminal and then try to reproduce the bug.
After you manage to get a button stuck, note the current time and then
try to stop the script with CTRL+C and upload the created button.log
file somewhere for review along with a note about the time the button
became stuck. If the problem is in the X layer, we should be able to see
it in that log.

Jason
---
Now instead of four in the eights place /
you’ve got three, ‘Cause you added one /
(That is to say, eight) to the two, /
But you can’t take seven from three, /
So you look at the sixty-fours....

On 06/10/2016 12:21 PM, Camille Bissuel wrote:

> Hi dear Wacom users list,
>
> In a desperate attempt, I'm reporting a nasty strange bug to your reading…
>
> I suppose you all now Krita 3 is released, and that's a nice release ;)
> Krita 3 is based on QT 5, and is supposed to have multi-touch support.
> I encounter a very nasty bug with Krita 3 : I get stuck in an always
> clicking state with QT5 apps, meaning when the bug occur, I'm clicking
> everywhere by simply moving the cursor hover any interface. So
> everything become unusable and I have to log off my session or reboot to
> temporarily resolve it.
>
> I already filled a lengthy bug report on Krita bug tracker :
> https://bugs.kde.org/show_bug.cgi?id=363225
> But as it's clearly an upstream bug now, and as Krita team is already
> overwhelmed, I'm looking for some guidance on how to solve that and
> being able to use my new shiny tablet with Krita again ;)
>
> I know there are very clever people on this list, and some that are
> particularly aware of the Linux input stack… I don't really know if the
> bug is coming from X server, from Wacom drivers or from QT5… and I'm not
> a developer… please help if you can !
>
> My best wishes, and many thanks for your time and attention in advance.
> Camille Bissuel - nylnook.com <http://nylnook.com>
>

------------------------------------------------------------------------------
What NetFlow Analyzer can do for you? Monitors network bandwidth and traffic
patterns at an interface-level. Reveals which users, apps, and protocols are
consuming the most bandwidth. Provides multi-vendor support for NetFlow,
J-Flow, sFlow and other flows. Make informed decisions using capacity
planning reports. https://ad.doubleclick.net/ddm/clk/305295220;132659582;e
_______________________________________________
Linuxwacom-discuss mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/linuxwacom-discuss
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Stuck in always "clicking" state with QT5 apps on a Wacom Cintiq 27QHD Touch

Camille Bissuel
Thanks a lot Jason, I knew you can bring lights to such darkness ;)

I managed to reproduce, and you can find the logs files here :
https://drive.google.com/folderview?id=0B-5J--XsvWh-Z3lWVksyVlJwMzA&usp=sharing

It seem the bug started when I grabbed the mouse at 18:33:21… but I'm not absolutely sure. It started in the 18:33 minute for sure.

Does it make any sense to you ?

My deepest gratitude !

Camille Bissuel - nylnook.com


2016-06-13 17:52 GMT+02:00 Jason Gerecke <[hidden email]>:
Camille,

It would be good to know if this issue is introduced in X or Qt so that
we can know what layer is broken. I believe the following command should
be able to straightforwardly show the press/release events occurring at
the X layer:

$ xinput test-xi2 --root | awk -v OFS='\t' '/Press|Release/ {X=$0;
getline; print strftime("%H:%M:%S"), X, $0; fflush()}' | tee button.log

Go ahead and run that in a terminal and then try to reproduce the bug.
After you manage to get a button stuck, note the current time and then
try to stop the script with CTRL+C and upload the created button.log
file somewhere for review along with a note about the time the button
became stuck. If the problem is in the X layer, we should be able to see
it in that log.

Jason
---
Now instead of four in the eights place /
you’ve got three, ‘Cause you added one /
(That is to say, eight) to the two, /
But you can’t take seven from three, /
So you look at the sixty-fours....

On 06/10/2016 12:21 PM, Camille Bissuel wrote:
> Hi dear Wacom users list,
>
> In a desperate attempt, I'm reporting a nasty strange bug to your reading…
>
> I suppose you all now Krita 3 is released, and that's a nice release ;)
> Krita 3 is based on QT 5, and is supposed to have multi-touch support.
> I encounter a very nasty bug with Krita 3 : I get stuck in an always
> clicking state with QT5 apps, meaning when the bug occur, I'm clicking
> everywhere by simply moving the cursor hover any interface. So
> everything become unusable and I have to log off my session or reboot to
> temporarily resolve it.
>
> I already filled a lengthy bug report on Krita bug tracker :
> https://bugs.kde.org/show_bug.cgi?id=363225
> But as it's clearly an upstream bug now, and as Krita team is already
> overwhelmed, I'm looking for some guidance on how to solve that and
> being able to use my new shiny tablet with Krita again ;)
>
> I know there are very clever people on this list, and some that are
> particularly aware of the Linux input stack… I don't really know if the
> bug is coming from X server, from Wacom drivers or from QT5… and I'm not
> a developer… please help if you can !
>
> My best wishes, and many thanks for your time and attention in advance.
> Camille Bissuel - nylnook.com <http://nylnook.com>
>


------------------------------------------------------------------------------
What NetFlow Analyzer can do for you? Monitors network bandwidth and traffic
patterns at an interface-level. Reveals which users, apps, and protocols are
consuming the most bandwidth. Provides multi-vendor support for NetFlow,
J-Flow, sFlow and other flows. Make informed decisions using capacity
planning reports. https://ad.doubleclick.net/ddm/clk/305295220;132659582;e
_______________________________________________
Linuxwacom-discuss mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/linuxwacom-discuss
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Stuck in always "clicking" state with QT5 apps on a Wacom Cintiq 27QHD Touch

Jason Gerecke
Maybe not so much light as you'd hoped... I don't see anything like a
smoking gun in the log output you provided, either around the 13:33:21
mark or anywhere else in the log. At first it looked like device #2
might be getting stuck, but on further investigation that's probably
just an artifact of running under GNOME -- my 24HDT produces similar log
output without any issues.

The only weird thing I see in the log is that at 18:26:12, a
RawButtonRelease is sent for "device: 2 (14)" (the touchscreen), but
there was never an associated press. Indeed, that one line is the *only*
time that device is ever mentioned. It would be difficult for a
release-without-press to cause a stuck button though, and the timing
doesn't match up with when you noticed an issue.

Based on the logs and the information you've provided in the bug, I'm
tempted to agree that this sounds like an issue with Qt rather than some
other layer. It almost sounds like Qt is dropping events from other
devices once the pen comes in proximity, and doesn't always properly
clean up their state by releasing any buttons that may have been pressed.

I've been able to reproduce the scenario you outlined in comment #8 with
my 24HDT, but no luck yet with the steps in comment #7.

Jason
---
Now instead of four in the eights place /
you’ve got three, ‘Cause you added one /
(That is to say, eight) to the two, /
But you can’t take seven from three, /
So you look at the sixty-fours....

On 06/13/2016 09:49 AM, Camille Bissuel wrote:

> Thanks a lot Jason, I knew you can bring lights to such darkness ;)
>
> I managed to reproduce, and you can find the logs files here :
> https://drive.google.com/folderview?id=0B-5J--XsvWh-Z3lWVksyVlJwMzA&usp=sharing
>
> It seem the bug started when I grabbed the mouse at 18:33:21… but I'm
> not absolutely sure. It started in the 18:33 minute for sure.
>
> Does it make any sense to you ?
>
> My deepest gratitude !
>
> Camille Bissuel - nylnook.com <http://nylnook.com>
>
> <http://nylnook.com>
>
> 2016-06-13 17:52 GMT+02:00 Jason Gerecke <[hidden email]
> <mailto:[hidden email]>>:
>
>     Camille,
>
>     It would be good to know if this issue is introduced in X or Qt so that
>     we can know what layer is broken. I believe the following command should
>     be able to straightforwardly show the press/release events occurring at
>     the X layer:
>
>     $ xinput test-xi2 --root | awk -v OFS='\t' '/Press|Release/ {X=$0;
>     getline; print strftime("%H:%M:%S"), X, $0; fflush()}' | tee button.log
>
>     Go ahead and run that in a terminal and then try to reproduce the bug.
>     After you manage to get a button stuck, note the current time and then
>     try to stop the script with CTRL+C and upload the created button.log
>     file somewhere for review along with a note about the time the button
>     became stuck. If the problem is in the X layer, we should be able to see
>     it in that log.
>
>     Jason
>     ---
>     Now instead of four in the eights place /
>     you’ve got three, ‘Cause you added one /
>     (That is to say, eight) to the two, /
>     But you can’t take seven from three, /
>     So you look at the sixty-fours....
>
>     On 06/10/2016 12:21 PM, Camille Bissuel wrote:
>     > Hi dear Wacom users list,
>     >
>     > In a desperate attempt, I'm reporting a nasty strange bug to your
>     reading…
>     >
>     > I suppose you all now Krita 3 is released, and that's a nice
>     release ;)
>     > Krita 3 is based on QT 5, and is supposed to have multi-touch support.
>     > I encounter a very nasty bug with Krita 3 : I get stuck in an always
>     > clicking state with QT5 apps, meaning when the bug occur, I'm clicking
>     > everywhere by simply moving the cursor hover any interface. So
>     > everything become unusable and I have to log off my session or
>     reboot to
>     > temporarily resolve it.
>     >
>     > I already filled a lengthy bug report on Krita bug tracker :
>     > https://bugs.kde.org/show_bug.cgi?id=363225
>     > But as it's clearly an upstream bug now, and as Krita team is already
>     > overwhelmed, I'm looking for some guidance on how to solve that and
>     > being able to use my new shiny tablet with Krita again ;)
>     >
>     > I know there are very clever people on this list, and some that are
>     > particularly aware of the Linux input stack… I don't really know
>     if the
>     > bug is coming from X server, from Wacom drivers or from QT5… and
>     I'm not
>     > a developer… please help if you can !
>     >
>     > My best wishes, and many thanks for your time and attention in
>     advance.
>     > Camille Bissuel - nylnook.com <http://nylnook.com>
>     <http://nylnook.com>
>     >
>
>

------------------------------------------------------------------------------
What NetFlow Analyzer can do for you? Monitors network bandwidth and traffic
patterns at an interface-level. Reveals which users, apps, and protocols are
consuming the most bandwidth. Provides multi-vendor support for NetFlow,
J-Flow, sFlow and other flows. Make informed decisions using capacity
planning reports. https://ad.doubleclick.net/ddm/clk/305295220;132659582;e
_______________________________________________
Linuxwacom-discuss mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/linuxwacom-discuss
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Stuck in always "clicking" state with QT5 apps on a Wacom Cintiq 27QHD Touch

Camille Bissuel
Thanks Jason,

You're right, that's not as much light as I hoped, but you reproduced the bug so I'm feeling less alone, and your test allow me to confidently fill a Qt bug report ;)
So that's a lot !

I will link it here as soon as it's ready…

Cheers,
Camille Bissuel - nylnook.com


2016-06-14 2:15 GMT+02:00 Jason Gerecke <[hidden email]>:
Maybe not so much light as you'd hoped... I don't see anything like a
smoking gun in the log output you provided, either around the 13:33:21
mark or anywhere else in the log. At first it looked like device #2
might be getting stuck, but on further investigation that's probably
just an artifact of running under GNOME -- my 24HDT produces similar log
output without any issues.

The only weird thing I see in the log is that at 18:26:12, a
RawButtonRelease is sent for "device: 2 (14)" (the touchscreen), but
there was never an associated press. Indeed, that one line is the *only*
time that device is ever mentioned. It would be difficult for a
release-without-press to cause a stuck button though, and the timing
doesn't match up with when you noticed an issue.

Based on the logs and the information you've provided in the bug, I'm
tempted to agree that this sounds like an issue with Qt rather than some
other layer. It almost sounds like Qt is dropping events from other
devices once the pen comes in proximity, and doesn't always properly
clean up their state by releasing any buttons that may have been pressed.

I've been able to reproduce the scenario you outlined in comment #8 with
my 24HDT, but no luck yet with the steps in comment #7.

Jason
---
Now instead of four in the eights place /
you’ve got three, ‘Cause you added one /
(That is to say, eight) to the two, /
But you can’t take seven from three, /
So you look at the sixty-fours....

On 06/13/2016 09:49 AM, Camille Bissuel wrote:
> Thanks a lot Jason, I knew you can bring lights to such darkness ;)
>
> I managed to reproduce, and you can find the logs files here :
> https://drive.google.com/folderview?id=0B-5J--XsvWh-Z3lWVksyVlJwMzA&usp=sharing
>
> It seem the bug started when I grabbed the mouse at 18:33:21… but I'm
> not absolutely sure. It started in the 18:33 minute for sure.
>
> Does it make any sense to you ?
>
> My deepest gratitude !
>
> Camille Bissuel - nylnook.com <http://nylnook.com>
>
> <http://nylnook.com>
>
> 2016-06-13 17:52 GMT+02:00 Jason Gerecke <[hidden email]
> <mailto:[hidden email]>>:
>
>     Camille,
>
>     It would be good to know if this issue is introduced in X or Qt so that
>     we can know what layer is broken. I believe the following command should
>     be able to straightforwardly show the press/release events occurring at
>     the X layer:
>
>     $ xinput test-xi2 --root | awk -v OFS='\t' '/Press|Release/ {X=$0;
>     getline; print strftime("%H:%M:%S"), X, $0; fflush()}' | tee button.log
>
>     Go ahead and run that in a terminal and then try to reproduce the bug.
>     After you manage to get a button stuck, note the current time and then
>     try to stop the script with CTRL+C and upload the created button.log
>     file somewhere for review along with a note about the time the button
>     became stuck. If the problem is in the X layer, we should be able to see
>     it in that log.
>
>     Jason
>     ---
>     Now instead of four in the eights place /
>     you’ve got three, ‘Cause you added one /
>     (That is to say, eight) to the two, /
>     But you can’t take seven from three, /
>     So you look at the sixty-fours....
>
>     On 06/10/2016 12:21 PM, Camille Bissuel wrote:
>     > Hi dear Wacom users list,
>     >
>     > In a desperate attempt, I'm reporting a nasty strange bug to your
>     reading…
>     >
>     > I suppose you all now Krita 3 is released, and that's a nice
>     release ;)
>     > Krita 3 is based on QT 5, and is supposed to have multi-touch support.
>     > I encounter a very nasty bug with Krita 3 : I get stuck in an always
>     > clicking state with QT5 apps, meaning when the bug occur, I'm clicking
>     > everywhere by simply moving the cursor hover any interface. So
>     > everything become unusable and I have to log off my session or
>     reboot to
>     > temporarily resolve it.
>     >
>     > I already filled a lengthy bug report on Krita bug tracker :
>     > https://bugs.kde.org/show_bug.cgi?id=363225
>     > But as it's clearly an upstream bug now, and as Krita team is already
>     > overwhelmed, I'm looking for some guidance on how to solve that and
>     > being able to use my new shiny tablet with Krita again ;)
>     >
>     > I know there are very clever people on this list, and some that are
>     > particularly aware of the Linux input stack… I don't really know
>     if the
>     > bug is coming from X server, from Wacom drivers or from QT5… and
>     I'm not
>     > a developer… please help if you can !
>     >
>     > My best wishes, and many thanks for your time and attention in
>     advance.
>     > Camille Bissuel - nylnook.com <http://nylnook.com>
>     <http://nylnook.com>
>     >
>
>


------------------------------------------------------------------------------
What NetFlow Analyzer can do for you? Monitors network bandwidth and traffic
patterns at an interface-level. Reveals which users, apps, and protocols are
consuming the most bandwidth. Provides multi-vendor support for NetFlow,
J-Flow, sFlow and other flows. Make informed decisions using capacity
planning reports. https://ad.doubleclick.net/ddm/clk/305295220;132659582;e
_______________________________________________
Linuxwacom-discuss mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/linuxwacom-discuss
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Stuck in always "clicking" state with QT5 apps on a Wacom Cintiq 27QHD Touch

Camille Bissuel
Just reported in the Qt bug tracker here : https://bugreports.qt.io/browse/QTBUG-54085
Please feel free to complete my report with any comment !

Thanks again for your work !

Camille Bissuel - nylnook.com

2016-06-14 9:10 GMT+02:00 Camille Bissuel <[hidden email]>:
Thanks Jason,

You're right, that's not as much light as I hoped, but you reproduced the bug so I'm feeling less alone, and your test allow me to confidently fill a Qt bug report ;)
So that's a lot !

I will link it here as soon as it's ready…

Cheers,
Camille Bissuel - nylnook.com


2016-06-14 2:15 GMT+02:00 Jason Gerecke <[hidden email]>:
Maybe not so much light as you'd hoped... I don't see anything like a
smoking gun in the log output you provided, either around the 13:33:21
mark or anywhere else in the log. At first it looked like device #2
might be getting stuck, but on further investigation that's probably
just an artifact of running under GNOME -- my 24HDT produces similar log
output without any issues.

The only weird thing I see in the log is that at 18:26:12, a
RawButtonRelease is sent for "device: 2 (14)" (the touchscreen), but
there was never an associated press. Indeed, that one line is the *only*
time that device is ever mentioned. It would be difficult for a
release-without-press to cause a stuck button though, and the timing
doesn't match up with when you noticed an issue.

Based on the logs and the information you've provided in the bug, I'm
tempted to agree that this sounds like an issue with Qt rather than some
other layer. It almost sounds like Qt is dropping events from other
devices once the pen comes in proximity, and doesn't always properly
clean up their state by releasing any buttons that may have been pressed.

I've been able to reproduce the scenario you outlined in comment #8 with
my 24HDT, but no luck yet with the steps in comment #7.

Jason
---
Now instead of four in the eights place /
you’ve got three, ‘Cause you added one /
(That is to say, eight) to the two, /
But you can’t take seven from three, /
So you look at the sixty-fours....

On 06/13/2016 09:49 AM, Camille Bissuel wrote:
> Thanks a lot Jason, I knew you can bring lights to such darkness ;)
>
> I managed to reproduce, and you can find the logs files here :
> https://drive.google.com/folderview?id=0B-5J--XsvWh-Z3lWVksyVlJwMzA&usp=sharing
>
> It seem the bug started when I grabbed the mouse at 18:33:21… but I'm
> not absolutely sure. It started in the 18:33 minute for sure.
>
> Does it make any sense to you ?
>
> My deepest gratitude !
>
> Camille Bissuel - nylnook.com <http://nylnook.com>
>
> <http://nylnook.com>
>
> 2016-06-13 17:52 GMT+02:00 Jason Gerecke <[hidden email]
> <mailto:[hidden email]>>:
>
>     Camille,
>
>     It would be good to know if this issue is introduced in X or Qt so that
>     we can know what layer is broken. I believe the following command should
>     be able to straightforwardly show the press/release events occurring at
>     the X layer:
>
>     $ xinput test-xi2 --root | awk -v OFS='\t' '/Press|Release/ {X=$0;
>     getline; print strftime("%H:%M:%S"), X, $0; fflush()}' | tee button.log
>
>     Go ahead and run that in a terminal and then try to reproduce the bug.
>     After you manage to get a button stuck, note the current time and then
>     try to stop the script with CTRL+C and upload the created button.log
>     file somewhere for review along with a note about the time the button
>     became stuck. If the problem is in the X layer, we should be able to see
>     it in that log.
>
>     Jason
>     ---
>     Now instead of four in the eights place /
>     you’ve got three, ‘Cause you added one /
>     (That is to say, eight) to the two, /
>     But you can’t take seven from three, /
>     So you look at the sixty-fours....
>
>     On 06/10/2016 12:21 PM, Camille Bissuel wrote:
>     > Hi dear Wacom users list,
>     >
>     > In a desperate attempt, I'm reporting a nasty strange bug to your
>     reading…
>     >
>     > I suppose you all now Krita 3 is released, and that's a nice
>     release ;)
>     > Krita 3 is based on QT 5, and is supposed to have multi-touch support.
>     > I encounter a very nasty bug with Krita 3 : I get stuck in an always
>     > clicking state with QT5 apps, meaning when the bug occur, I'm clicking
>     > everywhere by simply moving the cursor hover any interface. So
>     > everything become unusable and I have to log off my session or
>     reboot to
>     > temporarily resolve it.
>     >
>     > I already filled a lengthy bug report on Krita bug tracker :
>     > https://bugs.kde.org/show_bug.cgi?id=363225
>     > But as it's clearly an upstream bug now, and as Krita team is already
>     > overwhelmed, I'm looking for some guidance on how to solve that and
>     > being able to use my new shiny tablet with Krita again ;)
>     >
>     > I know there are very clever people on this list, and some that are
>     > particularly aware of the Linux input stack… I don't really know
>     if the
>     > bug is coming from X server, from Wacom drivers or from QT5… and
>     I'm not
>     > a developer… please help if you can !
>     >
>     > My best wishes, and many thanks for your time and attention in
>     advance.
>     > Camille Bissuel - nylnook.com <http://nylnook.com>
>     <http://nylnook.com>
>     >
>
>



------------------------------------------------------------------------------
What NetFlow Analyzer can do for you? Monitors network bandwidth and traffic
patterns at an interface-level. Reveals which users, apps, and protocols are
consuming the most bandwidth. Provides multi-vendor support for NetFlow,
J-Flow, sFlow and other flows. Make informed decisions using capacity
planning reports. https://ad.doubleclick.net/ddm/clk/305295220;132659582;e
_______________________________________________
Linuxwacom-discuss mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/linuxwacom-discuss
Loading...