HOW TO: Set up the HP TX2z and Dell XT & XT2 (N-trig digitizer) in Ubuntu

[Ubuntiac]

So, I'm a bit confused...

How does the announcement of Ubuntu's uTouch framework fit in with what's being worked on in this thread? Will it depend on the drivers / scripts being developed here? Would it be instead of these? Is it something completely unrelated?

[Ayuthia]

So, I'm a bit confused...

How does the announcement of Ubuntu's uTouch framework fit in with what's being worked on in this thread? Will it depend on the drivers / scripts being developed here? Would it be instead of these? Is it something completely unrelated?

From what I understand, the framework will allow our touchscreens to be able to use multitoutch gestures within Ubuntu. It will require the drivers that are discussed here but things have settled down a little bit with the kernel module so it could mean that we might not need an update to the kernel module when we use Maverick.

As far as I know, there are no new features to help with the rotation (and turning touch on and off) for evdev at this point so the scripts will still be helpful.

[Pikachuu]

Which means 10.10 will be a must-get for us?

[Favux]

Most likely. Especially if the 2.6.35 kernel has the 5-4-10 hid-ntrig.ko. Rafi could tell us. Then the only thing you should need to do is make sure you're running updated firmware. Preferably the 4.6.5.8.5 firmware in the 2.239 software bundle. Or newer if something better comes out. The caveat being that it's not totally clear to me that muti-touch will be implemented in Maverick, maybe just enabled. I think I read that full multi-touch is in Xserver 1.9, although maybe they'll backport it to Maverick's 1.8.

Is the touch toggle script working for you still?

[Ayuthia]

Based on all the information that has been flowing through this thread, I think this is how I have seen the devices listed using the most recent version of the hid-ntrig module:

Vista:
Can have two N-Trig Touchscreen devices so the one with the lowest id number from "xinput --list" is used. It reports the second finger as a distance from the first finger.

Win 7:
2.59 - Is reporting as two N-Trig Touchscreen devices. If I remember correctly, it might be the second report of the N-Trig Touchscreen that is used. If the second finger is placed on the screen, the data stops reporting. There is a kernel-module source that can get the multiple fingers to report, but in some cases it creates a lot of ghost touches (the cursor jumps around the screen randomly).

Note: The two reports of N-Trig Touchscreen/Pen is because in the earlier firmware versions, the data would be listed using if0 and if1 in the hardware reports. So when the kernel module is defining them, it sees them both and creates the devices even though one does not do anything.

2.187 (HP Version)- Reports as N-Trig Multitouch. It only reports two fingers.

2.172 (Dell Version)- Reports as N-Trig Multitouch. It reports up to four fingers.

2.239 - (HP and Dell) - Reports as N-Trig Multitouch. It reports up to four fingers.

If you remove the firmware - The hardware should default to single touch mode and it will report as N-Trig Touchscreen.

Please let me know if any of this is incorrect.

[Pikachuu]

I get phantom reactions when drumming all five fingers onto the screen at once with the touch supposedly off, but that's about it. The cursor doesn't even activate. Strangely enough the cursor always ends up close to where the thumb is. But by and large working. Thanks. Just to be clear, the script simply created an extra boolean in some system folder and checked that boolean to see if the touch was on or off?

EDIT: Slightly irrelevant, but as expected, since I made the changes to my script to disable touch on rotation to tablet mode, I found that that touch toggle script doesn't actually detect if the touch is attached, but rather detects through a boolean if it has been activated and reacts accordingly. Might want to note that so that people who are even newer at this than me don't get frustrated at the odd clashing behaviour.

If only I could get the media button to rotate the screen and then I'll be done; but I have work to do in preparation for the coming exams...

[Ayuthia]

Most likely. Especially if the 2.6.35 kernel has the 5-4-10 hid-ntrig.ko. Rafi could tell us. Then the only thing you should need to do is make sure you're running updated firmware. Preferably the 4.6.5.8.5 firmware in the 2.239 software bundle. Or newer if something better comes out. The caveat being that it's not totally clear to me that muti-touch will be implemented in Maverick, maybe just enabled. I think I read that full multi-touch is in Xserver 1.9, although maybe they'll backport it to Maverick's 1.8.

I agree with you, Favux. It looks like the code is enabled in Maverick, but I am not for sure if any (or how many) of the applications will be updated to use the code yet.

[Ayuthia]

I get phantom reactions when drumming all five fingers onto the screen at once with the touch supposedly off, but that's about it. The cursor doesn't even activate. Strangely enough the cursor always ends up close to where the thumb is. But by and large working. Thanks. Just to be clear, the script simply created an extra boolean in some system folder and checked that boolean to see if the touch was on or off?

EDIT: Slightly irrelevant, but as expected, since I made the changes to my script to disable touch on rotation to tablet mode, I found that that touch toggle script doesn't actually detect if the touch is attached, but rather detects through a boolean if it has been activated and reacts accordingly. Might want to note that so that people who are even newer at this than me don't get frustrated at the odd clashing behaviour.

If only I could get the media button to rotate the screen and then I'll be done; but I have work to do in preparation for the coming exams...

The script just checks to see if there is the temporary file out there. If it finds it, it will assume that touch is current off and will reattach the device back. If it does not find the file, it will assume that it is currently attached and will float the device ("turns off" touch) and create the file to say that it is currently off.

[Favux]

Hi Pikachuu,

Right. Putting a file in tmp is just a dodge to get around the fact that we don't know yet how to directly query the system as to whether evdev touch is off or on for "N-Trig Touchscreen".

[Ayuthia]

Hi Pikachuu,

Right. Putting a file in tmp is just a dodge to get around the fact that we don't know yet how to directly query the system as to whether evdev touch is off or on for "N-Trig Touchscreen".

Here is another way of doing it where it will check to see if the device is floating or not:

#!/bin/bash

# evdev touch toggle script

xinput --list|grep "N-Trig MultiTouch"|grep "floating slave" > /dev/null
TOUCH_ON=$?
if [ $TOUCH_ON -eq 1 ];then
    xinput float "N-Trig Touchscreen"
else
    xinput reattach "N-Trig Touchscreen" "Virtual core pointer"
fi

The xinput --list will list out the devices and the greps will look for a line that contains the N-Trig MultiTouch and floating slave words. If it finds a match, it means that touch is off. Otherwise touch is on.