Packaging a GTK app as an click package is possible, but I could only manage to do so by giving the app special permissions (so this is a workaround and will most likely not be excepted in the official app store). Here is an example that worked on my Nexus 4 with Mir as the display manager and will maybe also work on other devices:
Video of working app (a simple image editor):
https://libre.video/videos/watch/2f2c5304-d278-4b43-84e0-3c463d3a6432
Source Code:
https://gitlab.com/AdamSchrey/image_and_buttons

Some of the things done here will maybe not be necessary on some devices or in future versions of Ubuntu Touch.

I had to add this lines to the .desktop file:
X-Ubuntu-Touch=true
X-Ubuntu-XMir-Enable=true

I had to add this line to the .apparmor file:
"template": "unconfined",

Otherwise I would get this error:
** (process:26446): WARNING **: Unable to exec '/usr/lib/arm-linux-gnueabihf/ubuntu-app-launch/xmir-helper image-and-buttons_0.9.0 ut_session.sh' in '/opt/click.ubuntu.com/.click/users/phablet/image-and-buttons': Permission denied

I had to export this variable to get bigger buttons/text/entries...:
export GDK_SCALE=2

To get an on screen keyboard I installed this packages on my device:
maliit-inputcontext-gtk3 maliit-inputcontext-gtk2

...and exported this variable:
export GTK_IM_MODULE=maliitphablet

I reply to this old post, because when I searched for a solution, this was one of the first things that came up and I could not find any working examples my self.
Maybe someone will find this helpful or will tell me if my app and my solution also work on other devices.

Some of the things I did, were taken from this website:
http://kriscode.blogspot.com/2016/09/x-applications-on-ubuntu-phone.html

Update:
I removed the proof of concept web app from the OpenStore, because I have no time right now to regularly restart the server or to fix the stability problems. It would also not make much sense to improve something, that will not be used.

It is all available under GNU GPL3 at: https://gitlab.com/AdamSchrey/open-store-payment-poc
So if anyone wants to revisit the idea, he/she can do so.
I am also happy to answer any questions about it.

Thank you all for the feedback

@doniks I thought the icon could be improved and created a different icon:

This is a png file. The original file is a svg file. It seems like the forum does not like svg files, so I made a merge request.
I did not like the low resolution of the current icon, this should not be a problem with a svg file.
If you still prefer the current icon, that is no problem.

Hello,
I wrote an app to compare the performance of different devices and released it on the OpenStore: https://open-store.io/app/ut-sysbench-qt-gui
I hope the arm64 version works, because I can only test the armhf version.
It would be nice, if some of you could share your device's name together with the results you get, after clicking on "Multiple Benchmarks".
"Export and Compare" will copy the results to your clipboard and show you a bar chart with the performace of other devices.

Multiple results from different persons for the same model are also useful.

You are also free to share problems with my app. I know that the memory (RAM) test does not work on all devices.

----------------------------------------------------------------------------------------
Results of my LG Nexus 4 with Ubuntu 16.04 OTA-19:

sysbench --test=cpu --num-threads=2 --cpu-max-prime=20000 run
57 events per second

sysbench --test=cpu --num-threads=8 --cpu-max-prime=20000 run
113 events per second

sysbench --test=memory --num-threads=4 run
Operations performed: 0 ( 0.00 ops/sec)
0.00 MB transferred (0.00 MB/sec)

sysbench --test=mutex --num-threads=64 run
total time: 2.4057s

@Appointed
About UT and desktop apps:
The goal of Ubuntu Touch is to be an secure and open source mobile OS, that can optionally also turn into a Desktop OS, when you connect the mobile device to a bigger screen, mouse and keyboard.
When the user is in the desktop mode of UT, he maybe wants to use classic Desktop apps and this is what Libertine is for.
So it is not the goal of UT to make any desktop app work on a smaller formfactor and without a keyboard and mouse. Of cause you can still try to do so, but that it is not scaling correctly is no surprise. The real issue are the other errors (some programs not opening or not working correctly).

@Appointed said in Usable phone display:

responsible

Libertine is not working perfectly, but UT is a community driven project. You and me are as responsible to make things better as everyone else. The code is all on GitLab and GitHub. The UT core team is happy for everyone who can help.

For a lot of people it was way more important to get Android apps working on UT and move to focal and this was accomplished. So Libertine not working perfectly is a result of prioritizement and lack of contributors and not because of responsibility.

@MV I don't have a Volla X23, but the scaling probably works similar on most devices.
The steps in the question are maybe already the solution for you:

https://forums.ubports.com/topic/9317/how-can-i-change-scaling-on-20-04-focal-without-ut-tweak-tool-and-without-reboot

... These devices (and maybe other devices) set the scaling in /etc/deviceinfo/devices/halium.yaml.

In this file I changed "GridUnit" to a more convenient value (i.e. 28 on FP3 and 18 on Lenovo).

To make these changes take effect it is not sufficient to just restart Lomiri with systemctl --user restart lomiri-full-greeter.service. You'll have to reboot the device...

@kugiigi said in OpenStore payment system discussion:

Anyway, my concern right now is what's in the poc app? I'm not accusing anyone but we are mostly people who met in the internet and what if people who tries the poc actually input their payment details? How secure is the site and what does it actually do? Sorry if this was a stupid concern. It's just the first thing that came to mind

Your concerns are very valid, but it is designed so that the user does not need to put any data from him/her into the app. The user will see an address and can use any Litecoin wallet (another program or website) to send something to this address.
The server for open-store-payment-poc will now check the block chain, if any payment went to the given address (the sender and receiver are unknown, but every transfer to every address is known to everyone, this is how the block chain works).

You will always get a new address, when you click on "Buy" (needs some time and maybe sometimes fails, because the server hardware is not the best and my code for this proof of concept could be improved, it should only show that it can work). But every address always points to the same wallet (in this case my wallet, but it could also be the wallet of someone else for some other app).

@kugiigi said in OpenStore payment system discussion:

There's just so many things involved and not just technical but also legal matters.

The main idea here is: The OpenStore would not need to process the the payment, because it is done by multiple third parties (people who mine Litecoin or who have Litecoin nodes) and the OpenStore only provides an address and looks if there is anything transferred to that address. But the developer would probably need to take care of legal matters as taxes when he receives Litecoin to his wallet or later transfers them into "normal" money.

And just to be clear: Litecoin was chosen here as one of many possible solutions, because it works and is well established. I could have also chosen Bitcoin, but the confirmation times and the fees are lower with Litecoin.

@doniks You're welcome and no problem

Here is a similar thread for CPU performance tests with sysbench: https://forums.ubports.com/topic/6750/ut-sysbench-qt-gui-compare-cpu-performance-of-different-devices

Result of the graphic test of my LG Nexus 4:

@Salah I saw your other post and the "hacky" way with remounting the file system is the way I would do it (https://forums.ubports.com/post/81981).
I recommend to back up your data and to enable adb shell access, before doing so (you can often still access your device over adb shell, even if the GUI is broken).
Please be aware that an OTA update could break your device, after you did some hacks, but it is probably unlikely after changing some icons.
I don't think UT/Lomiri looks for any icons in the home folder (but I could be wrong), so it is either the "hacky" way or someone adds code to the system (pull request on GitLab), that lets Lomiri look for icon packages in a folder like $HOME/.icons/ (similar to what gnome-tweaks does).

Update:
I removed the proof of concept web app from the OpenStore, because I have no time right now to regularly restart the server or to fix the stability problems. It would also not make much sense to improve something, that will not be used.

It is all available under GNU GPL3 at: https://gitlab.com/AdamSchrey/open-store-payment-poc
So if anyone wants to revisit the idea, he/she can do so.
I am also happy to answer any questions about it.

Thank you all for the feedback

@Fuseteam Yes, you are right. The transaction is directly between the wallets from the user and the developer, but it would maybe be good to implement a system that ensures that the user can ask the developer for a refund (for example a button that appears after the payment).
Maybe it would even be easier to have a system where the user can use the app for free for a certain amount of time (the time he would be entitled to a refund) and than the user can decide to pay for the app, if he wants to get future updates.

This is of cause very theoretical, because with the feedback I got so far, I don't think the majority of users want any form of payment system.

@arubislander It should not be necessary to pay again. The user will receive a confirmation code, that will also be saved on the server. So later on the user can use the confirmation code to download the app again.
Additionally, if the user has an OpenStore-Account (with Ubuntu One, GitLab or GitHub), it could maybe be even more convenient for the user (just logging in again).

In the proof of concept, the confirmation code is always 1234, but in a real scenario it would be a new code for every generated address.
To prevent the confirmation code from getting lost, it could be automatically stored on the device and maybe (if the user wants this) be send to an e-mail address.

Of cause if you have a friend who already payed for the app, you could just use a confirmation code from him/her and it will probably not attract attention, unless the confirmation code was already used 300 times this month.
For this I want to loosely quote the additional information text from the proof of concept: "I don't think preventing people from getting apps for free should be the future goal of the OpenStore. Apps (that require payment) will be purchased, when the payment is the most convenient way for the user."

@kugiigi
Sorry for all the information, it was just important for me that the user does not need to give any risky data to the app.

@kugiigi

This is how the screen should look like (with some other random address of many addresses that all point to the same wallet), when you click on "Buy".
The green arrows point to the amount that was send to the given address.
As soon as the server detects, that the cost of the app was send to the given address (Confirmed address balance >= Cost of app), you will get a green "Install" and "Download" button.
You could also try it out by getting the code from gitlab and run it locally or you try it out with 0.0000001 LTC (I think you can make the amount you send as small as you want, but I only tried it with amounts between 0.0001 LTC and 0.001 LTC).

@kugiigi said in OpenStore payment system discussion:

Anyway, my concern right now is what's in the poc app? I'm not accusing anyone but we are mostly people who met in the internet and what if people who tries the poc actually input their payment details? How secure is the site and what does it actually do? Sorry if this was a stupid concern. It's just the first thing that came to mind

Your concerns are very valid, but it is designed so that the user does not need to put any data from him/her into the app. The user will see an address and can use any Litecoin wallet (another program or website) to send something to this address.
The server for open-store-payment-poc will now check the block chain, if any payment went to the given address (the sender and receiver are unknown, but every transfer to every address is known to everyone, this is how the block chain works).

You will always get a new address, when you click on "Buy" (needs some time and maybe sometimes fails, because the server hardware is not the best and my code for this proof of concept could be improved, it should only show that it can work). But every address always points to the same wallet (in this case my wallet, but it could also be the wallet of someone else for some other app).

@kugiigi said in OpenStore payment system discussion:

There's just so many things involved and not just technical but also legal matters.

The main idea here is: The OpenStore would not need to process the the payment, because it is done by multiple third parties (people who mine Litecoin or who have Litecoin nodes) and the OpenStore only provides an address and looks if there is anything transferred to that address. But the developer would probably need to take care of legal matters as taxes when he receives Litecoin to his wallet or later transfers them into "normal" money.

And just to be clear: Litecoin was chosen here as one of many possible solutions, because it works and is well established. I could have also chosen Bitcoin, but the confirmation times and the fees are lower with Litecoin.

@Moem said in OpenStore payment system discussion:

@johndoe That would completely stop me from using apps from the Open Store. I don't use any payment methods on my phone. I also do not use cryptocoins of any kind, and do not wish to.

So there is no form of payment method for the OpenStore that would be acceptable for you or is it just about the solution I mentioned?

@Lakotaubp Oh okay, I thought Ubports would be involved with the OpenStore, because it comes with the OS and seems like the official store of UbuntuTouch. Is the OpenStore not running on servers from the Ubports foundation?

Hello,

I know that the OpenStore shows donation links for some apps, but this is not quiet the same as selling the apps. I think it would be beneficial for UbuntuTouch, if developers could decide to take a small amount of money for their app.
Developers probably cannot make a living from selling UT apps at the moment (small user base). But even a small amount for an app can be very motivating to improve that app and the performance of the OS as a whole. Improved apps and an improved operating system can attract more users to pay and donate, and that has the potential for even more improvements.

This is easier said than done and there are many things that would need to be thought about.
This is why I made a small proof of concept under: https://open-store.io/app/open-store-payment-poc
https://gitlab.com/AdamSchrey/open-store-payment-poc (no time to fix stability problems of the server, so people who want to try it out have to look at the code and run it locally)

I think one possible solution for the payment system would be to make it with Litecoin. If we use a blockchain-based cryptocurrency, we have the inconvenience for the user to exchange the "normal" currency into a cryptocurrency. But the maintainers of the OpenStore would have the advantage of not having to process the payment and only providing an address that corresponds to the app and check for transactions to that address.

I would love to hear your thoughts on this topic. I hope the discussion can center around whether this is beneficial for UbuntuTouch and if the approach over blockchain-based cryptocurrency is feasible.

@Salah
The others are competly right about the potential risks.
But I do not want to discourage you from hacking the device, if you wish to do so and if you know that it could maybe break some things.

Put this in the terminal, if you want to have fun hacking and potentialy breaking your device:

sudo mount -o rw,remount /