Ubuntu with phosh?

Is this possible? Could I run Ubuntu on the phone using phosh as the DE?

Edit: Let me flesh this out a little better.

I know that Ubuntu has an official build for Raspberry Pi, which means that the heavy lifting of getting Ubuntu setup for ARM is already done. Of course there is fine tuning to worry about, and whether or not the drivers necessary to use the L5’s hardware are available and functional.

But I am assuming all of this is possible, and that if Phosh works, you should be able to use pretty much whatever Debian based Linux distro you’d like.

Would this be correct?

Ubuntu has packages for phosh, phoc, squeekboard, etc., but you don’t have an installer, so I doubt that you will get the right u-boot files and Linux device tree.

You could install PureOS on the Librem 5, then switch your /etc/apt/sources.list to Ubuntu repos, and then do an apt upgrade and see what happens. It will probably break something, but it is always fun to see how much you can screw up, before you are forced to do a clean reinstall. :wink:

Considering that there isn’t an Ubuntu port for the PinePhone, I doubt that the Librem 5 will get one either. However, there is a Mobian port for the Librem 5. Anything that you like about Ubuntu, you can probably figure out how to install and/or configure in PureOS and Mobian with a little work.

1 Like

Oh, I’m not so hung up on the idea. It would be cool getting snaps working on the L5 though. I really like the idea behind snaps. (which is that there is a need for a central repository that simplifies getting software for newcomers, etc.)

Thanks for the reply.

A quick look in the PureOS repo hints that you can get snaps working much easier than by installing a new distro. Btw, any reason you’d prefer snaps over flatpaks (which Purims intends as default portable format)?
I could imagine there are more non-free apps in the snap store, but possibly not in ARM versions :slight_smile: !?

I just had a look at this article and thought somebody might find it interesting:

1 Like

Well i can say it isn’t because of any technical specification. It really is because I believe duplication of effort and fragmentation are big problems for Linux. I see what the snap store is trying to do and I think it has the best chance of actually accomplishing what the people behind it set out to accomplish with it.

You are probably right about the arm issue with snaps. However getting something like Canoncial into this for the Librem 5 would be a huge boon to the end user. Canonical has not made any secrets about their own efforts to push into the mobile arena. Like this their efforts wouldn’t be so largely wasted. Although that might only be semantics.

I was under the impression that Canonical gave up on mobile in 2017 when it abandoned Ubuntu Touch. Has anything changed since then?

1 Like

those could be side-loaded in the similar way as I’m doing it for the L5 image on pinephone. The main challenge I think will be in pre-configuring the rpi-ubuntu to work on mobile form-factor (be it l5 or pp).
I’d imagine it will be something like Arch image for PinePhone which just gives you CLI terminal and then you’re on your own to bring in phosh and make it work.

Take the bootloader and kernel from PureOS and put those into debootstrapped Ubuntu arm64 rootfs. Tada, you’re done. You should be able to install pretty much any arm64 distro this way (provided that it has recent enough components like Mesa, otherwise you’ll run into troubles).

I believe there was a video on Hackersgame YouTube channel about running OpenSuSE on the Librem 5 this way.


Evidence that the Librem 5 vision is unfolding just as envisioned! Thanks @dos!


And we’ll call it “Phobuntu!” :wink:

Isn’t it ironic that Canonical with all of its clout couldn’t do what Purism has done with the L5? Furthermore that mobile Ubuntu is possible today because of what Purism has done? Makes me wonder where the problem really was.

At any rate, who cares. Thanks Purism and keep on keeping on!


No it’s not ironic. It’s the different mindset. Canonical ended the project to turn to other things that could be more profitable. A lot of the technology that Canonical develops is said to be caused by the not-invented-here syndrome. They duplicate something that already exists, then after some years give up on it and use what everybody else uses. (e.g. Mir/Wayland)
Canonical tried to create their own mobile eco system, controlled by them, with their own desktop stack (Mir/Unity instead of Wayland or Xorg). Trying to mimic Google/Android with their own eco system. Yet another incompatible thing that developers should support.
Purism on the other hand forked an existing Wayland compositor, which might in turn become the Gnome default when Xorg is sunset. They contributed to Gnome, adding mobile capabilities, such that it was adopted by many distros before the L5 was even shipped. They succeeded because making quick profit was not the motivating factor (even though many don’t get that, looking at the price of the L5)


I think you aren’t seeing the whole picture. Canonical is thought to be greedy, but I’d say they are just realistic. It takes money to pay developers to see quality work that is sustainable. Unity was because Gnome was a mess. Canonical doesn’t have to sit and wait to see what the community will drum up. It has a team of developers who are all competent and can churn out work.
I would argue that everything you see as wasted effort was actually Ubuntu simply trying to live up to what it is trying to accomplish. They want the end user experience to be consistent and stable. If the tools that Debian are pushing forward aren’t quite there, they then go and build something themselves. As they open source all of it, it is rather naïve to believe that the work they do isn’t benefiting software elsewhere.
Both Purism and Canonical are under the same constraints: Money. Purism couldn’t wax long about their love of FOSS ideals if they weren’t making money selling hardware. Ubuntu couldn’t sustain its development, likewise, if it wasn’t making money in other ways.
The beef over snaps is irrelevant. New users who come to Linux are turned off the second they need to use software not in a default repo. Snaps are a solution, and I haven’t seen anything else across the sphere that is trying to do the same thing.

I mean the user experience is paramount. Just today, after adding a swap partition back, I started getting the GRUB menu popping up. No explanation, no reason for it. A few searches later, and an obscure command found with grub updated, the situation was remedied. My mom being able to do this, is just not realistic. On top of that, WHY did it even happen. Stuff like this happens all the time. If you add an extension to Gnome and it doesn’t work, there is not an intuitive way to see what went wrong. It is not even that clear how to remove the non-working extension… No matter how capable Linux is today, there are glaring holes in its preparation.


In my opinion, Canonical, Mozilla and Jolla failed at Linux phones, because they tried to just produce the software, and then rely on hardware companies to market Linux phones. The problem is that the phone makers were not fully committed to Linux and didn’t market the phones to the right group of consumers who could appreciate Linux’s unique properties.

Another thing to keep in mind is that the cost of developing a mobile phone in Shenzhen has dropped dramatically since the time when Canonical thought that it would need to raise $32 million in preorders to bring the Edge to market, and Jolla ran into financial problems trying to ship hardware. Both companies gave up on creating their own hardware, and tried to just sell the software, and they were never able to find hardware companies that marketed their phone properly to the right audience.

Another problem was that both Canonical and Mozilla tried to create a lot of their operating system from scratch, whereas Purism has tried to design Phosh as a thin-overlay on top of the existing Wayland+wlroots+GTK+GNOME stack to minimize software development and long-term maintenance costs.


All companies, software developers, designers and usability experts focus on Phosh only now and have all devices be able to run the same OS with ultimate GUI, all apps in a single app store.

Next, apply the Nexus strategy by having a pool of manufacturers showing off what they are capable of by pushing the Phosh devices to highest levels in terms of technology, marketing and awareness/acceptance of the market.

Manufacturers will be shipping out their devices with Phosh before you know it. They have been constrained by Google for too long, so have million of users. Next step for the manufacturers is to make the devices blobless and we should be able to see a major shift in the market.

It will even make DRM companies (Netflix, Spotify etc of this world) aware they need to push their platforms towards RYF in order to maintain their position.

After that fork whatever you want, just make sure the open source is presented as a unified product with excellent basics. From that point forward educate the users what they can do more, so much more with their Linux devices.

Then people will have the choice to what they really want and move away from the Big Techs. In my opinion this is the only way to make open source a success and generate revenues to keep all idealistic parties alive.

1 Like

I don’t think it’s helpful to portray past efforts as failures, partly because each of those efforts continue in some form (UBports, KaiOS, SailfishOS) but also because it establishes a narrative of failure around mobile Linux.

In addition, some of the marketing for Firefox OS was done by operators like Telenor, who sold (or planned to sell) phones based on Firefox OS in a number of markets, like Serbia, Hungary and Montenegro. There was quite a push for developers during that time. Getting operators on board with promotion and outreach has been shown to be important when establishing a new mobile platform, though it remains to be seen if that is still the case.


@2disbetter The following metapackages are also available from Debian Testing:

phosh-core (just the basics)
phosh-full (Everything: games, calls and chatty, other stuff)

So I would image that either they already are available in Ubuntu, or will be soon.


You mean Mobian for Pinephone right? No installer for other phones.

1 Like

Thanks! I think it is largely superficial to run Ubuntu on the phone over PureOS. You can simply point to ubuntu repos and you’d be 90% of the way there.

I disagree, no one in the Linux world wants to be locked into one interface. I have no desire to run phosh and don’t even want it on my phone, I’ll patiently wait until plasma mobile is usable enough to run without crashing.