Ubuntu Touch Q&A 55

Edging Forward
30 июля 2019

 
 

Looking for the Audio-only version of the Q&A?

You're just one big orange button away -->

News and Updates

Mir

Some of the problems we have had with the new adaptations to UT are caused by apps that do not function properly with Mir. There are conflicts with work done earlier by Canonical. The approach we have taken now is that if it does not work as we want, we kill the app. iOS and Android also kill apps in similar situations, so this is pretty normal practice. It avoids problems that follow when apps don’t behave correctly. The inability to close apps while not in focus has now been fixed.

There needs to be a major refactoring in qtMir and Unity8. It general though, apps take a very complicated route when handled by Mir, so it is not easy to solve problems and the refresh needed will have to be big.

Nevertheless, with those two changes we are very close to the point where Edge channel can be merged into Development channel.

TELEports

Florian announced a couple of weeks ago that a big new version of TELEports is on the way but QA has showed that there are still some snags. The plan is still to release as soon as possible. Of course with holiday season, that will take a bit of extra time to achieve.

In the new release forwarding will work, with an additional comment box. So, annotated messages. When forwarded from another channel it will be tagged with the channel it came from. We need a bit more work on multi-select for messages still. Multiple deletion will then be possible too. There will also be some special message types, for contacts and location. Vcard support is more complicated and will come later.

Also included are lots of translation updates and even some extra languages.

A QML function called backflux is very useful in this connection. But QML code does need to be kept very clean otherwise it slows everything down and causes problems.

OTA-10

OTA-10 will bring much better GPS in practice, by removing the assistance functions. It will be less sophisticated but as a result more reliable. Rebooting retains the previous location, so the fix is quite fast. There is now documentation available about the use of location services.

‘How to use’ is what we want to introduce across the board. We also need to do that for deep things like Mir and Unity8. Alberto has already documented Calendar in a lot of detail. That should be our aim for all of the core apps. Taking notes as you learn is a great approach to writing documentation.

As a side note, many of the big advances in development are made in the middle of the night but that is absolutely not recommended as a way of getting things done!

Community Gratitude

Aurelio has submitted 50 Yumi images which have now been converted to .png and from there into stickers for Telegram. Anyone in the supergroup can now download the pack.

Sponsors were thanked.

Marius lost his connection part way through the Q&A because he was using mobile data for HD video!

A reminder that questions can be posted in the News section of the forum.

Questions

Is the move from Github to Giltalb ongoing?

The answer is that all core apps have been moved over and we are continuing to move stuff. All the system deb packages are still on Github and built on the Jenkins server. That infrastructure is very complicated and consequently much more difficult to move. Added to that, the people who are competent to do it are our core developers, so it means diverting their efforts away from the OS. So a balanced approach is needed.

Gitlab is actually very nice in terms of organising stuff rather than just coding. We have gold status as a community, so we enjoy a form of indirect sponsorship on Gitlab. On there, we build clicks, which are automatically pushed to the OpenStore. That is very useful and saves us a lot of time. All of our OTA-10 progress is tracked on Gitlab.

How is the progress with PINE64 devkits?

Netkit has been working a lot on the PINE64. It uses a chipset for graphics which is new to us. It needs a version of GNUutils later than that found in our kernel. Lima kernel is still glitchy though. We might find that the only solution for now is to use the proprietary mali driver to get it to work. All of this needs to be built into Edge to make sure it works. It would be a big mistake to try to build those changes and Edge in parallel, hoping to merge later. One of the basic principles of UT is that there should be no visible difference in the look and functioning of the OS, whether it is working on an Android phone, on PINE or on anything else.

Will 18.04 ever be built?

Will we use systemd on 18.04? The answer is that we don't know yet. Upstart has the big advantage for now that it works on our legacy devices. The closed source on older device trees blocks the use of systemd. There is no benefit anyway while we are in 16.04. Logically, we will probably jump directly from 16.04 to 20.04 because the effort will be about the same but the longer period of support will better justify the work necessary.

Are we still talking to companies about making a phone with UT?

With PINE64 and Purism the answer is yes. With others, although we have had some contact, their obstacles to working with us are much higher. If they major on Android phones they will want very substantial reassurance before taking us on. PINE64 could be the practical example that convinces them. They will not want to get stuck with software warranty etc.

PINE64 have the fewest obstacles to supporting us and we help them by providing a ready-made OS. No non-disclosure agreements etc are required. PineTabs are on the way now incidentally. It is impressive that while PINE have lots of different projects on the go at once, they are delivering on all of them.

Is there any news about Librem 5?

Well, two devkits have arrived. Marius got his just this week. No more news than that yet, obviously. We have to prioritise and for us PINE64 comes first but the work on that will directly inform what we do for Librem, so it is not really a case of first one, then the other. The cpu is different in each. That will be the main challenge.

What is the status of Unity8 on the desktop and are developments are being fed into Edge channel?

While all of the modifications being made in Edge should end up in the desktop branch, the primary focus of our community is on the phone. Desktop matters to us but it will never be our top priority. There is a clear gap in the market for a team that wants to specialise in Unity8 for desktop. They would be welcomed by us and they would not be weighed down by the phone issues which we have to contend with. Just a reminder - ‘convergence’ means ‘one codebase running everywhere’. It does not mean – as many think - one device running everything!

Is the idea still to move towards benefiting from upstream wherever possible?

What elements are still missing from that plan? Ofono is the single biggest example of something that we still maintain ourselves. Strictly speaking, Intel would be the real upstream for that but Yallow in Sailfish is where we would actually go. locationd is the other canonical element that we have to think about. NetworkManager is at least on the way to upstream. It still needs some patches.

Slightly different but Cordova support existed once but has since stopped. It would be great if we could use that again.