News and Update
OTA-13 little delayed
OTA-13 is delayed a little by a critical bug. Text messages are not waking the device long enough to be seen. It is now expected in Stable next week. There are Stable updates for Xperia X and other Xperias. The Morph browser has been improved a lot, including grab handles for text selection.
Debian and Manjaro Movement
Qtmir will go into Debian.
There is an image link for anyone who wants to try it. Manjaro will be using Arch packages rather than Clicks. Maybe snaps. Only core apps such as dialer are being played with at the moment. These are the early days of what will be a long project.
TELEports new feature
Sponsors were thanked.
The News section of our Forum is the best place to post questions for the Q&A. YouTube live chat, Telegram and Matrix are other places to post a question.
If you didn't know, the Forum questions get priority.
Fakone asked about Godot games in UT
We have no contact with them and they would need to help us if we wanted to enable their games on UT. It would need someone with a foot in both communities, ideally. They already have packs for Snaps and Flatpaks so in theory it might be possible to use those as the basis for a click template. It would be a big undertaking though and definitely not a project for the core team.
Alter asked for clarification of a discussion last time round on the merits and practicalities of including a firewall in UT. His understanding was that there is in fact a firewall, which you can choose to turn on in terminal. There was brief mention of possible problems caused by older kernels. Could that be explained more? Well yes, terminal activation is possible but the question was partly about whether to have it active by default and if not whether to have a user friendly way to operate it selectively.
Older kernels do not contain all the APIs that newer software use. The firewall software is moderately old, so the that reduces the challenge a bit. However, each device would need its own adaptation and that certainly would add to the complexity of creating a functioning system for all. The other thing is that those older kernels have not been tested for that purpose and rigorous testing would be needed. There are likely to be bugs which were not fixed by manufacturers, in parts of the kernel which they never needed to use.
Android doesn’t use a firewall. It does have a system which we could maybe adapt. But the other consideration is that those old kernels are being dropped anyway, as they move to mainline.
UT qemu documentation
Gizmochicken asked about UT qemu and any documentation for it. Well there is no documentation. It is something that Marius put together, really for his own use. It does function to some extent. It is quite hacky and is intended to get UT working in a virtual machine. It is a simple, short bash script. A snap version was made at one point and put in Edge but it was pretty unstable and would need a lot more work to sort out. Maybe some day it could be developed further.
Webapp push notifications
Rocket2nfinity asked whether webapp push notifications will come with the qtwebengine upgrade? The answer is no. Dalton has been reading up about webpush and found that (unexpectedly) it is quite like UT push. The webapp server is told to send a message to a webpush gateway. Your device then connects to the webpush gateway. Even if we could figure a way to add notification functionality to Morph, it would still only work when the browser is in the foreground, which rather defeats the point of notifications. The similarity does maybe point however to some way of integrating webpush technologies with UT at some time in the future.
What about redshift and videochat?
Is anything happening with those now that have a new qtwebengine? Redshift implementation would happen in the mir layer, so is not really impacted by the qtwebengine change. The way that cameras work on Android also mean that there is nothing in this upgrade which will automatically give us videochat capability.
Marius played around with red filters in Lomiri and the effect was not at all good. Creating colour shifts in Mir worked a bit better but not much. That is the right way to go though. A new composer (Android hwcomposer service) implements a way to do it so maybe a solution is half built for us. That would be limited to newer versions of Android though. Swapping colour values live using cpu is very heavy on resources so we would have to avoid that sort of ‘mechanical’ solution.
Ubuntoutou asked about Florians toy plane kit. Which Florian showed. It is a project for ‘later’ :)
Why was Morph based on chromium and not Firefox?
Why not use something like Firefox Focus? The answer is very simple. Qt chose chromium and we chose Qt. The long answer is that browsers are vast projects. Even maintaining a browser is way beyond our capabilities and resources. Think of them these days as an OS in their own right.
Chromium’s embedding is really easy to do, whereas for Firefox it is quite complicated. A way of providing hooks between chromium and gecko has been tried but it is a very hacky approach. Servo was a very interesting project but it has been terminated? Hopefully some sort of gecko based approach will develop at some time in the future.
Aury88 had a question about why Sphinx was chosen for our documentation, rather than something like MediaWiki? Actually, moving to Sphinx rather than a wiki was one of the first things that Dalton organised when he joined the project. There used to be a wiki and it was vandalised endlessly, creating a load of work. Sphinx has version control built in and while it isn’t easy, it is easy to manage.
Dalton’s original plan was that the docs should provide a full guide to using the phone, aimed at the ordinary user. As it turned out, other tasks were always more pressing and we never got to that point.
Actually there are niches within the UBports documentation where we are using wikis. The Android 9 documentation is an example of that. Hopefully though those will shift over to Sphinx over time. It has a format which is easier to read and it can generate e-books. Being able to read documentation offline is very useful if you are trying to concentrate. Incidentally, Readthedocs hosts for us. Yet another point is that Sphinx has full support for our translations interface.
It was asked if there is any more progress with the PinePhone? Although it might not seem relevant, the work Marius is doing with Manjaro has a lot to do with this because they are working, for example, with a newer version of ofono. They also have the newer kernel which we need to upgrade to. Getting Lomiri out to a wider audience will draw in development efforts from further afield, which will also directly benefit UT. So it is a cooperative project which assembles greater resources than we have alone, to accelerate some of the core work which we must do. Testing should not be forgotten either. By testing wider and in a greater variety of settings, we can have a big impact on stability. We have jumped from Qt 5.9 to 5.12, which is a really big undertaking. More integration with other projects will mean that we can more easily and smoothly go just a step at a time.
Essentially all of our components are in a similar situation to Qt in terms of versions and progress, so joint working will benefit almost every part of our OS. It makes us a little sharper and tidier about our code, knowing that others will have to rely on it too. A bug in history-component had lain undiscovered until Marius went through and did some tidying up and annotating. It didn’t return anything with return…
UT and Manjaro are polar opposites of the software spectrum. UT relies a lot on patching old components derived from Canonical whereas Manjaro is based on Arch, with components from the bleeding edge. Pulling those two together so that they work is really quite an amazing thing to do.
If you fix warnings, the code is more optimised by the compiler and you get speed improvements. The cumulative effect is that things start to fly. Marius and Dalton were very bubbly about the project alongside Manjaro because it is actually fun and low stress. That is a refreshing change and the importance of it should not be underestimated. The big difference is that development is not weighed down by the packaging tasks. Manjaro have a light and elegant approach to packaging, leaving developers to develop. Small detail that it ignores tests… but that is another story :)
Will UT run on the Pixel 5?
The short answer is no. The slightly longer answer is that it isn’t even released yet… We can confirm that Google have not approached us to put UT on the Pixel 5. Seriously though, we are still stuck on Android 9 so 11 is quite a way off. A reminder though that there is a port for the Pixel 3a. It uses the GSI from Erfan and therefore needs a manual install.
UT coolest feature?
It was asked asked what the coolest feature is for getting a newbie hooked on UT? The simple answer is of course convergence. Privacy is frankly more important but it is maybe more obscure from a newbie perspective. The independence of UT and the fact of the community around it are potentially attractive features though. Stability is a good selling point, along with having a very frequent update cycle. We are now no longer entirely alone in convergence of course, as Windows surface has something similar.
Everyone will have their own reasons for choosing UT and their priorities may not necessarily be the same as those of ‘people like us’.
Developers have a particular angle on using the device as they know in detail what goes into it. An average user is much better placed to give their impressions to other users.
See you next time :-)