@shano said in Why is Wayland compositor and Lomiri so far in terms of functionality compared to Compiz and Unity7 released some 20 years ago?:

Compiz was released in 2006 and is still maintained. Unity7 was released in 2010 and is still maintained. My question is why Wayland and Lomiri are so far away in terms of functionality, plugins etc. compared to tech made 20 years ago.

Are we building on top or reinventing the wheel?

From what I see, Lomiri is not reinventing Unity 8, it is Unity 8.
I downloaded the Lomiri source code and counted roughly 18000 commits between 2013 and 2017 (end of Ubuntu involvement) and 1400 commits between 2018 and 2025. The main reason is probably that Unity was a business project with serious resources behind it, while Lomiri (the new name for Unity 8 since 2020) is an open source project with a few part time volunteers. Also, it may be that there was so much work in Unity that there is not a lot of new development necessary, the bulk of new work being in other parts of the full (and huge) smart phone stack. I did not count but I think that a large part of these 1400 commits are just translations.

I don't know Compiz but I tracked the project in its last hideout on Gitlab and counted about 30 commits in last 5 years, 20 of them in 2020. It's maintained in the most limited sense, there is almost no development.

Here are the videos recorded about mobile:

https://video.fosdem.org/2026/ub4132/

UT is mentioned in the videos about state of Foss on mobile and the one about the apps store: the UT one is rather well placed, however Flatpak is coming for mobile and it has already a dominant position in general Linux apps.

The video about openHarmony is intriguing given the resources of the organisations behind it.

I have been interested by the one about push notifications, apparently more ready for prime time that I was thinking. It seems that the only real solution for Linux clients is the kde one, that would seem a good fit for UT since it uses QT.

@Sander said in Add "Tab" key to Ubuntu Touch keyboard?:

You can tap anywhere on the terminal area to mimick the behavior of the tab key

thanks for this hidden gem ! TIL

Duh ! I will close this stoopid post, I had not used seriously a smartphone since several years and did not realize anymore how sensitive to location Gps services are. I tried only in 2 places in my home and both where unsuitable. When I got out in the air, Gps started to work. So it is working, really. I did a stroll this morning and it updated its position. I have still to get my hands to another brand of smartphone on Android to compare with the Fairphone 5 under UT to see how well it is working.

@poVoq

hello, not the dev here - but the links are not broken for me. The app is supposed to be based on

https://github.com/luigi311/ConverseJS-ubports

the links are a bit strange though (why not a software forge ? if the author don't like Github, there are other such as codeberg...). It's not stated in the original repo that the author has given up. Personally I'd not use this software without more information.

As an update, a draft MR was recently posted on Gitlab using the BeaconDB api (directly,not via Geoclue)

@uxes said in Enabling MAC randomization:

shipped on our system by default

I am not sure that any phone is doing that by default.
It has also a downside for anyone using this phone with ssh, that is, the IP address affected by the Dhcp server (the wifi access point) will change often.
It's not a big deal but it can be annoying.

@GooglyBear said in GPS don't seem to work on FP5 UT 24.04 stable:

a GitLab issue to follow progress for Fairphone 5

AGPS is not really linked to a specific port (even if it can work on some devices it seems, but it's mostly unintended). It's tracked here

@Averiz

Congrats ! this is a great idea !

I tested it on my FP5 (24.04-1.1 stable) and it installed and restored the system keyboard. Needed to restart the phone but that's not important. Tested it with the terminal and a few browsers.

Good points:

1. I find it more pleasing to the eye that the system one.
2. It worked as well with alternate browsers (uWolf, Min) who have special methods to handle the keyboard.
3. with the terminal it does not interfer with the special input methods (double clic on the screen, special area

Minor points:

1. opening Min (an alternate browser) and Morph both with a blank initial screen, entering 'https:' and clicking on the blank screen generated a mad flickering of the keyboard and a Js error in the browser code (I took a screenshot but for some reason it's unreadable).

2. when setting / unsetting it as system keyboard there is some delay after clicking the button with no apparent feedback.

3. I (and I think most people) expected it to respect the system preferences for the language, nope, it's necessary to setup the keyboard in its own config screen. Yes, it's actually a high expection This config screen is not totally obvious by the way. I managed after a few mistakes to select the French Azerty as the default and it displayed finally - in Qwerty . I clicked on the 'AZ' key and it was Azerty finally.

4. after restoring the system OSK I set it again as the default, and I got back the qwerty with the french Azerty left selected, and this time clicking on AZ did not change the disposition, it stayed Qwerty.

5. still with the French Azerty, the text 'Keyboard - French - French (AZER) spills over the key ',' at the left (see screenshot)

@kristatos said in Smooth Edges (name pending) - Let's Fix the Bugs That Drive You Mad:

if the input is to slow?

FIY I have noticed that disabling haptic feedback was leading to me typing faster. I don't know if it's psychological or if the buzz is really slowing the keyboard. Give it a test if you don't have disabled it yet.

@_Joao_

for all 'modern' sites you should use Morph QT6 instead of Morph. Get it from the store.

@domubpkm

Well, if my calculation is correct that means about 2Mbits/s, it seems quite decent for Bluetooth.

gpatel-fr said:

3.1.1

BTW this does not look normal: the version is not incremented while there is about 500 LOC added in the RC version.

BTW 2: I have done already 3 back and forth between 1.2 and 1.3 RC2 and I have trouble reproducing the problem, that is, I have seen the problem in my last reinstall of 1.2 . Color me baffled

@peat_psuwit said:

Unlike system components and some system apps, Gallery app is a Click application where updates are shipped from OpenStore.

color me surprised.

phablet@ubuntu-phablet:~$ ls /usr/share/click/preinstalled/gallery.ubports/
3.1.1  current

gpatel-fr said:

I was not able to use the Recorder

cancel that, it had absolutely nothing to do with the upgrade, it was just a problem in my phone configuration. Recorder works fine.

This morning after having tested that the new version gallery problem was indeed 'fixed' by returning to the previous stable, I wanted to upgrade to RC and again I see the lack of upgrade channels. For screenshot see https://forums.ubports.com/post/95155 . This may be a new problem, or not. I have never wanted to switch back and forth like that previously.

Edit: I restarted the phone again and the channels were back to normal. This is a transient problem.

Edit 2: checking again, the Update parameters screen is displaying 'Channel: candidate', the system state displays '1.2' and checking updates says that everything is up-to-date. This is a bit messed up

Edit 3: after 2 hours of trying of updating, failing, restarting, switching channel, restarting, failing, I finally discovered in a flash of insight into the minds of advanced developers that the update process is actually testing the update history, the things that are displayed below 'recent updates', and this case of getting RC and stable back and forth has never been really thought to be possible (after all, regressions are something that have never been seen for more than one time or two in all the eons of software development), and so unlocking the switch to RC again needed to delete the previous update history.
So I'm finally on RC3 again. Oups, edit 4: RC2

Looking at the log with my FP5 back to stable 24.04-1.2 I see no such messages.
Messages about 'binding loop' whose fixing was one of the goal of the change.
I see one message about property of nulll but it's of 'isSelected' instead of 'show' and 'path'.

@Linus67

well, there is a choice of error messages about undefined properties in gallery code.

After looking at the MRs, there is one (1) change since 2 years. I don't see anything else but it's not a small change, rather a collection of logically unrelated changes.

@peat_psuwit
should not a 'stable' branch (something like 24-04-1.x that IIRC exist in some other repos) be in the repo and changes be backported only after having seen some testing in the 'unstable' branch ?

@Linus67

thanks, this is so clear that I reinstalled 24.04-1.2 stable and it works fine indeed !
That's a clear regression. Now I'll go to gitlab look at the MR for the gallery app.