Live discussion at: xmpp:livewire@conference.agnos.is?join

For the past ~2 months, I have been hacking away at making a native XMPP client for Ubuntu Touch. When @brenno.almeida released Greenline, the opportunity was too good to pass up. There exists an XMPP client called Libervia (formerly known as Salut a Toi) that works as a background daemon, and thus could fit pretty much straight into the existing architecture of Greenline.

Many experiments later, we now have Greenline but yellow the internals have been completely redone in an unholy abomination of Rust (of course) and Python (because Libervia is written in Python). It's a fully confined app that supports most modern XMPP basics, with a few features that still need to be implemented (e.g. reactions).

The client is considered to be in alpha state, but is available on the OpenStore for testing. I consider it functional enough for daily use, but with rough edges. This thread is a discussion/support thread for the app. More specific bug reports should be filed at the git repository: https://git.agnos.is/projectmoon/livewire

You can find it on the OpenStore. I expect a handful of updates in the next days/weeks to address the most glaring bugs, and then to work on proper push notification support, which will be its own interesting journey.

Indra is now available on the OpenStore!

https://next.open-store.io/app/indra.projectmoon/

This thread will serve as a sort of unofficial support thread for now. Proper issues/bugs/feature requests should be done on the issue tracker.

Indra is a Ubuntu Touch app for off-grid communication using LoRa (long range) mesh radios. You can text friends, family, and neighbors without an internet connection. Messages are sent via the radio, which connects to other radios nearby to form a network completely separate from traditional communication infrastructure.

You need an external radio to use this app. They are cheap to build or buy, and a one-time investment in off-grid communication. Indra currently supports radios running Meshtastic firmware. You can connect to your radio over Bluetooth, the local network, or a USB cable (depending on your radio).

The app currently does not support configuring or setting up your radio; please use the Android/iOS app or the Meshtastic web client at https://client.meshtastic.org to set up your radio or change its settings.

This is a work-in-progress app. The initial release supports what the author considers the essential, core features of Meshtastic. Currently supported features:

• Connect to a configured radio over Bluetooth, Wi-Fi, or USB.
• Send messages to public channels, private channels, or directly to other users on the mesh.
• View and search the list of other radios that your radio is picking up.

Things currently NOT supported:

• Configuring the radio. Please use the Android/iOS app or Meshtastic web client to set up the radio.
• Showing positions of radios on the map (next major feature to be implemented).

@bgriffis check KeePassRX on the Open Store.

Disclaimer: I am the creator of KeePassRX.

Documentation: https://agnos.is/projects/keepassrx/

I've taken the liberty of packaging Lagrange for Ubuntu Touch. Lagrange uses its own completely custom C-based UI stack built on top of SDL/OpenGL, which means that while it runs fine, the keyboard does not work out of the box on Ubuntu Touch. The UT package has a patch that 1) enables the keyboard and 2) packages it as a .click.

There are still some issues with the keyboard and integration with the system. Namely:

1. Clipboard does not work.
2. Keyboard still has some weird behavior on textboxes that aren't the URL bar. You cannot edit text in the middle of them, for example.
3. No integration with content hub.
4. No integration with URL dispatcher (i.e. cannot open gemini links in Lagrange).

I think most of these should be solvable. I will continue working on it and see what comes out.

You can find Lagrange here: https://next.open-store.io/app/fi.skyjake.lagrange/

After seeing the Hydroxide app on OpenRepos.net for Sailfish OS, I was inspired to port it to Ubuntu Touch. A few crazy ideas later, and it's now a Rust app that runs the official Proton Mail Bridge instead of Hydroxide (alternative bridge).

So, this means you can now connect directly to Proton Mail with Dekko, or any other e-mail client you manage to get running on your UT device.

Find it in the Open Store: https://next.open-store.io/app/projectmoon.hydronium/

This thread will serve as a basic support/questions thing. For more detailed bug reports. please file an issue in the repository. I am particularly interested in any problems with the bridge itself, especially around the untested secondary mailbox password login.

@kugiigi @pparent I think snaps should be turned on by default, but only ones verified to work properly with Ubuntu Touch. Then a setting to turn on showing results from all snaps, if user desires. Of course, this requires someone to go curate snaps.... So in the mean time, maybe changing how the results are displayed + opt-in setting is better.

@domubpkm please don't just dump AI output into the forum.

@undrwater also, since this is not built in, you could check out @kugiigi's Ambot Installer and install MariKit. It does what you want. Slide left or right from middle-ish of the bottom edge of the screen and it's gestures ahoy.

https://next.open-store.io/app/indra.projectmoon/

Indra is now available on the OpenStore, with all the fixes in Alpha 4 and a few more tweaks! From this point forward, the database will be treated as something that needs to be migrated instead of wiped on every update. (The setting in the menu to wipe the database does not actually do anything, and will be removed).

Indra 1.2.1 is now available with a laundry list of improvements and fixes. Find it at https://next.open-store.io/app/indra.projectmoon/ or https://git.agnos.is/projectmoon/indra/releases/tag/v1.2.1.

• Feature: Make links clickable in messages.
• Feature: Add the ability to search and filter nodes on the map.
• Feature: De-duplicate messages that are likely rebroadcasts.
• Feature: Show hops away/signal strength on messages.
• Feature: Add/remove favorite nodes.
• Improvement: New message footer layout.
• Improvement: Better node list appearance in dark mode.
• Improvement: Tweaks to the choice of black or white text for messages in dark mode. More readable.
• Improvement: Add node ID to the node info page.
• Improvement: Show node distance on the map, and sort by distance.
• Improvement: New map icon.
• Improvement: Pull to refresh the node list.
• Improvement: Proper status indicators for sending messages.
• Improvement: Show channel/conversation status indicators on main view.
• Improvement: Add a setting to toggle forced Bluetooth reset.
• Improvement: Add a setting to control how long it takes before a node is considered offline.
• Improvement: Add a setting to control the message de-duplication window.
• Improvement: Translation updates.
• Fix: Apply the default channel name to the public channel no matter its index.
• Fix: Fix a UI lockup when reconnecting to the radio.
• Fix: No more UI flickering when disconnecting from the radio.
• Fix: More efficient refreshing of the node list.

@mihael said:

There is also another new issue (regression): when I am notified through the notification that a new message arrived, there is a significant delay until that message is displayed in the chat - both when that specific chat is opened or when the list with all chats is displayed.

I can confirm this (sort of?). It seems like the app, if it's already open, will not necessarily load the conversation that the notification points to. And when it does, it seems to not load the latest messages in the conversation until you go out and back in to the conversation.

@oldbutndy said:

@projectmoon said:

@domubpkm please don't just dump AI output into the forum.

So, @projectmoon, since AI answers not allowed (which I understand, because half of what I get when asking Gemini anything lately is contradicted BY GEMINI after I rephrase the question, and it is NOT allowed to learn from its own deeper dives !), please summarize the RCS implementation problem using no AI answers.
Also, please estimate a minimum timeframe from Ubuntu Touch to have a working implemtation.
I am not asking for a 'when will it be working' type answer.
I am asking for a 'it probably WON'T be working for, at absolute minimum, nn years' type answer. '
I am sincere in asking for this, because I don't find anyone anywhere suggesting it can be done.
I also found a human discussion that said major US carriers tried implementing their own versions, then gave up & let Google support it.
So, if it really is a thing that Ubuntu Touch won't support for years, it would seem to make sense to explore alternatives.
And support the best one strongly.

I didn't say don't use AI answers, I said don't dump raw walls of text straight from an AI chatbot because no one is going to read that. I have no idea what the timeframe for RCS on Ubuntu Touch is (non-existent at the moment, I would imagine). But when posting, one should at least expend the minimum effort to write their own words, even if using an LLM to get an answer. It's not hard.

@domubpkm please don't just dump AI output into the forum.

Is it even possible to implement RCS in a completely open source way at the moment? Since it seems to be stuck inside Play Services, and Apple just doesn't care about it? Like, could I go out and buy some random Nokia feature phone for 80 EUR and would it work with RCS?

RCS was supposed to be a replacement for SMS, which means it should Just Work (TM) at the SIM level after the implementation is done in the code for the thing that manages the SIM (ofono, in this case). But it seems like in reality it has been implemented in some weird way where Google somehow controls it entirely on Android, and not all RCS connections play nicely together? The whole thing is a mess.

@woodyst Love the concept, and there's definitely a need for more thorough GPS. But yeah, it needs translating. Is there a place to contribute translations? Or maybe you can generate a set of machine translations for the most used languages on UT and then people can update those? Would help people actually be able to use the app. My Spanish is way too rusty for this lol.

As for what I can see so far:

• Does not fit UT's/Lomiri's UX, which isn't necessarily a bad thing, but a more native appearance would look better in my opinion.
• It's unclear what the account is for. I assume traffic reports, and I assume it's probably explained in the tour I didn't read because I don't know enough Spanish lol.
• Might be nice, if this is possible (not sure it is) to integrate with system links for directions and stuff, e.g. responding to geo: URIs if you don't already.

@pparent said:

@projectmoon

So no bug in the end?

Currently, it seems not.

@kugiigi @pparent I think snaps should be turned on by default, but only ones verified to work properly with Ubuntu Touch. Then a setting to turn on showing results from all snaps, if user desires. Of course, this requires someone to go curate snaps.... So in the mean time, maybe changing how the results are displayed + opt-in setting is better.

@pparent said:

I've released V8.15.0-1 with the following changes:

-Fix keyboard closing after sending a message
-Bump to Signal-Desktop 8.15.0-1 (First to officially support arm64)

Edit: never mind, it seems to have settled now. This might have been due to openstore updating the app.

Behold! A crash on startup:

Jun 23 13:21:26 ubuntu-phablet systemd[9539]: Started lomiri-app-launch--application-click--signalut.pparent_signallauncher_8.15.0-1--.service.
Jun 23 13:21:26 ubuntu-phablet aa-exec[239595]: rm Error : '/home/phablet/.local/share/signalut.pparent/recently-used.xbel' does not exist.

@DerekZoZo said:

@projectmoon unlocked bootloader was not a problem whatsoever on Fairphone 6 when it comes to BankID in my personal experience.

Well that's good. Depends on the app, of course.

@Ida_ said:

@DerekZoZo I've tried, but not succeeded. Waydroid is still seen as a rooted system by the app, so it refuses to run. But maybe you find a way in the future. If you ever do, I would really appreciate a guide if you have the time.

I am getting some banking apps to work in Waydroid, so it is still very useful.

It probably is not "rooted," as such, but unlocked bootloader. That is, unless you actually rooted the Android image. it's still dumb, and I'm not sure unlocked bootloader can be "fixed" without locking the bootloader at the hardware level (which you should not do with Ubuntu Touch).