@mariogrip

FP5 build 144

I can half confirm @shafiqalibhai report about external storage, for me the external storage is seen and mounted (it can be used from the command line), however it is indeed not seen in Places.

For reference, here is how it appears on 24.04-1.3:

clicking on the 'name' (well, id) of the device allows to browse.

With 26.04, the display is identical but the entry marked with an arrow in my screenshot is missing.

While I'm at it, I have noticed that in Morph browser, it's not possible to share an URL: the menu entry (Share) exists, however clicking on it does nothing so it's not possible to create a web site icon with Webber (this works in 24.04-1.3)

@CiberSheep

the OONI report points that the decision of betting all on Cloudflare has downsides. Maybe use a mirroring strategy with another CDN ?

@PierreM

J'ai acheté un Fairphone 5 il y a maintenant 10 mois et je ne regrette pas l'achat, c'est un téléphone bien conçu et assez bien supporté par Ubuntu Touch.

Mais si j'achetais un téléphone maintenant je prendrais plus en compte l'alternative Volla.

Je n'avais pas retenu les téléphones non compatibles 5G et ça excluait tous les téléphones Volla sauf le Quintus. Celui-ci n'avait pas de batterie amovible et pour moi c'était aussi quelque chose d'essentiel.

Le Plinius a une batterie amovible et il fait la 5G donc il répond à mes besoins essentiels et il est fourni par une société qui supporte beaucoup plus UT que Fairphone.

Le point faible du Plinius est sa connexion USB limitée à USB 2 mais c'est pareil pour le FP6.

@Farmebrown

the trouble with using a talking electric heater as an advocacy helper is that it is making your thoughts look likely without proper basis.

In the case of your post, it is asking to Ubuntu Touch to provide an alternative to a Windows software called 'Windows phone link'. Why is this inappropriate ? because it is a desktop software. Ubuntu Touch is a phone software. Ubuntu Touch do not aim at being an alternative to Fedora, Debian Ubuntu...

If KDE Connect find this functionality out of scôpe, if no well known software phone exists on Linux that can do that with an Android phone, that's for a reason: it's because it's a marginal need for a market that is - if one takes 3% of the 1 billion users of desktop OS - about 30 million users.

If and when Linux desktop developers write such a software that can connect to an Android phone to do something equivalent to what you are asking, now it could be the moment to study why an UT phone could not be used instead of an Android phone.

I'd say that this hypothetical problem could probably be solved then entirely at the linux kernel level without any Ubuntu Touch application software. I think that the Ms software relies probably only on 'standard' Bluetooth features.

@mariogrip

FP5 build 137

Now it's the camera app that don't work anymore.
Like with the store, it's a missing library problem: libexiv2.so.27 called from libcamera-qml.so

26.04 provides libexiv2-28, but contrary to libicu74 there is a compatibility layer in 26.04: libexiv2-27-compat, but it's not installed.

I don't see why this breakage appears only now. Was this libexiv2-27-compat library included previously in the rootfs and not anymore ? I am not keen on the idea of a regression test by installing a previous version of Ubuntu Touch 26.04 to check on that.

Anyway, getting the library from the compat package and doing an overlay to add it to the system solves this particular problem.

gpatel-fr said:

libicudata.so.74

looking a bit more at this issue, this comes from the operating system upgrade: Ubuntu 24.04 comes with libicu74 while 26.04 uses libicu78.

Ubuntu 26.04 don't have libicu74, so don't seem to be able to use some binaries built for a previous Ubuntu version.

The use of snap in Ubuntu is something of a workaround for this problem I guess while click format was designed before the rise of snap/flatpak with their library packages.

@PhAndersson said:

Could it somehow be related to the fact that the clock is now reasonably synched?

I don't think so, for the following reasons:

• I did give you the hint about recreating the connection, and I don't have a mediatek device
• date problems for ssl come from the securing through a certificate infrastructure, that don't make much sense for Bluetooth, whose hacking radius is 10 meters vs 40000 km for standard internet networking
• updating certificates and certificates lists would be unfeasible for many Bluetooth devices that don't even have an Internet link

This said, if you want to try it out, it's easy: just re-enable the qualcomm time service and check again.

as noted in latest Q&A, in 26.04 alternate browsers running in XWayland are now displayed as any other app in the launcher, no more 'X' icon.

@kristatos

yes however logging happens at the generic journal level, there is no prefix to filter on. Entries are listed under 'aa-exec' and usually include the app name (ratatoskr) in the listed paths, you can try

journalctl -b | grep ratato

There is also logging done by obex that you can list using

systemctl --user -e -u obex

@kristatos

maybe doing some more tests.
I tried it again with my phone (FP5) setup with 24.04-1.3 stable and succeeded with receiving and sending test files.

However I have reproduced some of your problems.

Not the obex service not known. This was not happening for me.

I have seen the disconnection of Bluetooth happening though. Strangely the Bluetooth page for the device (an Android phone) was displaying the device as connected BUT with no signal, then disconnected very soon after (may seem logical given there was no signal).
I removed each phone from the Bluetooth configuration of both phones, tried again and this time the signal was rated 'excellent'. It then stayed connected. Needless to say, for all these tests the phones were sitting on a table very near of each other. It looks as if the state of connection is rated with 'no signal', it's staying so even if the real connection is good. In this case the pairing had been done much before and the phones had both been off or way to distant of each other to talk with Bluetooth.

After solving this connection problem, I think another trouble may be that if a transfer is not acknowledged by the receiving end, all following transfers fail. Not sure but it looks like that. This problem was happening with a transfer UT -> Android.
If my impression is correct, it may be a problem with the app (the connection problem mentioned before is probably more caused by the system bluetooth stack).

@dl8aax

my quick and mostly uninformed guess would be that for an app to do something intrusive like changing the sound level for the system, some authorization would (should) be needed.

So maybe tracking what is happening at the D-Bus level by starting 'gdbus monitor' just before initiating the change in your app and looking at the saved result could be a way to confirm (or not) this idea.

If that's the problem, it could be a matter of setting the right Apparmor policy for your app.

as noted elsewhere for 24.04-2, unstable now can provide the OSK to XWayland apps such as Chromium-UT, Min, UWolf.

@DerekZoZo

right, you have to know that first, the Plinius port is very recent and not well known outside of core devs, and second, core developers presence here is very limited.

The bulk of help here is provided by 'advanced' users, not developers.

@DerekZoZo said:

it felt like an attack the way you phrased that

@keneda said actually 'Maybe you should try first'

As for me I'll call this a 'suggestion', certainly not an 'attack'.

I'll share some reservations about the english speaking Volla forums (did not try the german ones). I did get an answer, but it was not very informative, so finally I answered my own question (much later).

Another suggestion (feel free to characterize that as an attack if you want) could be to post a work item on the Gitlab repo for the device here

You'll be more likely to interact directly with the person(s) doing the actual porting job. There is actually a work item created already. In it the dev seems to say it's possible with 'shenanigans', maybe could give you some tips if you ask nicely.

@pparent

since VoLTE is basically voice over data, could this have a relationship to the losses of data connectivity that you report also ?

@PhAndersson

Maybe you could try to disable the Timekeeper service ?

@PhAndersson

you are correct, I missed that because I'm mostly on 26.02 these days. I switched back briefly to 24.04 stable and I could repro.
It's probably a service order start problem since starting obex manually succeeds, I will have to investigate it.

@PhAndersson

then run the following command in Terminal app or via ADB/SSH:

gsettings set io.ubuntu-touch.system-settings.update offer-development-release true

@Vlad-Nirky

ah yes I see it now. By luck in this position it does not block the indicators.
Theoretically the software should elegantly adapt to all the quirks of the hardware. Practically it's not as easy as one could think

@Vlad-Nirky

I have seen already another post about things being messed up in landscape and staying messed up when back to portrait - something I definitely don't see on my FP5, where the landscape mode is not perfect but when switching back to portrait all is well.

I think I see the difference now: your phone has no notch for the back camera. It's probably the root cause. The MR has been tested for a phone like the FP5, not like your phone.