I would like to spitball a feature request. Last time I tried to put paid podcasts onto my UT device, I had two options: either send them from my laptop via USB or the like, or, use Firefox within Libertine to open the site where they were downloaded from (because last I checked Morph Browser wasn't capable of logging into the site in question, or rendering it, I forget which) to download them to the SD card. The site in question is Podbean, but there are other platforms which allow for premium subscriptions I think. A third option would be to run the Podbean .apk within Anbox and hack-sync the downloads from the Anbox filesystem to the UT filesystem, but ehh, I don't want such a crude brute force method. In any event, these three options would require the podcasts to be played in the Music app, but as the apps stand now, I prefer playing podcasts in Podbird to Music.

So, could we find a way to input credentials into Podbird and have it access specific archives of premium podcasts depending on which ones you are subscribed to? I believe Podbird syncs it's free podcasts according to the iTunes database, so can we implement this for other podcast databases which are password protected?

I think this will be my next UT device Two things I'm curious about are:

Some sources online state that the max SD card capacity is 256GB, other say 512GB. Has anyone tested for the upper limit?

Is the USB C port USB 3.1 rated? Or just slowmo USB 2?

Sorry I went quiet for a while. Thank you for your advice. I've seen a couple of drawbacks with Premiumize in the VPN and otherwise, such as lack of tsl-auth certification as you say, so I might consider moving to Nord when my subscription runs out. But for now I will continue to use Premiumize. Thankfully I was finally able to achieve a VPN connection on my Pro 5 by using nmcli to take the .ovpn file as input along with my username and password, and this also added the connection to the GUI elements so I can turn it on and off with the button. Just what I wanted. I had to do an apt install of the network-manager-openvpn-gnome package for this to work; as VPN editor cannot take .ovpn files at this time, I am looking into raising an issue with Ubports to see if it would be worth including this package in the system image by default. In case this this approach is of interest to anyone else struggling to add a VPN to Ubuntu Touch, the full command used was sudo nmcli connection import type openvpn file "yourfilename.ovpn". I think I had to run it twice for online IP checkers to actually confirm the phone as (supposedly) being in Montreal as was to be expected.

However, unfortunately now I have another problem. For personal reasons I have decided to keep my M10 on Vivid OTA-3 until such time as I feel the Xenial build is mature enough to upgrade to. But this does mean it may be vulnerable to bugs which Xenial doesn't have. I wonder if I have discovered one in attempting to apply the VPN to the M10 in the same fashion as for my Pro 5. Here, after running the aforementioned command, it responds Error: failed to load VPN plugin: missing "plugin" setting. Is anyone familiar with this error? I have searched for it online but can't find the string really anywhere except in the Network Manager source code itself, and there, the code doesn't really help explain what the problem may be at least not to my eyes. It does seem that whether this is a bug or something that can be fixed by installing a package or similar, this situation changed from Vivid 15.04 to Xenial 16.04. Can anyone shed some light? Perhaps it might help if you happen to have your own UT device still on 15.04, or a livecd image of 15.04 to test this on x86 architecture. I had considered just copying the VPN connection file from /etc/NetworkManager/system-connections from the Pro 5 to the M10, but I thought this wouldn't really make sense because for example if they shared the same uuid (which I presume is generated during VPN configuration) as a consequence, then they can't both be on the same WIFI connection simultaneously.

@Code_Gaug Your English seems fine to me on first glance my friend

Thank you for your pictures when they are uploaded. Your help is appreciated.

Thanks, I will try tinkering with that as soon as I have a moment. Meanwhile, I would still appreciate it if anyone can spot anything glaringly wrong with the setup I've tried to go with.

Here are the screenshots:

Note that on this second one the file in the ~/.cert/nm-openvpn directory was initially of .pem type rather than .crt type. I renamed the end of the file to .crt to see if that would help, but alas, still nothing.

If it would help, I could also include screenshots of the OpenVPN settings on Ubuntu 18.04 for a 1:1 comparison. If anyone would like to see them, please ask.

@Code_Gaug On seeing your post I was relieved because I thought this might allow me for the first time over several attempts ever since my first use of Ubuntu Touch to use my Premiumize VPN connection. But unfortunately it doesn't, not even enabling LCO data compression will help. In sheer frustration and desperation, I am resorting to sharing my OVPN file for one of the connections to see if anyone can help guide me into doing this. The inability to import the file into the UT VPN editor is proving a hindrance, despite my best efforts to import the particulars by manually choosing the settings. Is there something about the way Premiumize configures their VPN settings which precludes them from being used as a VPN with Ubuntu Touch? I don't know. Try to help me out here please, I don't know how to fix it and I need some advice. A step by step guide to the settings given the information I am about to provide would be optimal; I will soon follow up with another post to include some screenshots of the UT VPN editor with the particulars (e.g. username, password) blurred out.

I'll preface this by saying that the certificate has been completely scrambled for the sake of sharing this, so that nobody else can use the certificate for their own means but they can only help troubleshoot. I believe this is the only uniquely identifying information in the .ovpn file but please point out for me if I'm mistaken. I have used both the .ovpn file and the equivalent settings in Ubuntu 18.04 x86's VPN utility which were set when importing the .ovpn file for that system to try to set it up for Ubuntu Touch. Here is the text content of the file:

remote vpn-ca.premiumize.me
verify-x509-name CN=vpn-ca.premiumize.me
auth-user-pass
client
dev tun
proto udp
cipher AES-256-CBC
resolv-retry infinite
nobind
persist-key
persist-tun
mute-replay-warnings
<ca>
-----BEGIN CERTIFICATE-----
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
-----END CERTIFICATE-----
</ca>
verb 3
reneg-sec 0

@doniks Yes I know it's not a silver bullet, but that's why both times I mentioned it, it was alongside "other improvements" which will give the full experience. For example I should think the segfaults mentioned above might need other fixes. And yep, the run of an individual developer in this community can indeed be transient in nature.

Steven, I appreciate you couching your words to avoid offence, but I disagree on your points.

@TotalSonic said :

There is not really any available device that would benefit from using the outdated and unsupported UT OS, and it would be a serious waste of time to any of the UT developers to try and accommodate this as well.

You seem to be implying that I suggest that the developers to go back to the old way of doing things, that is to support Vivid which is a version several years out of date (and which was so even up to the point of UBP OTA-3), which would be akin to keeping a zombie on a life-support drip. That is not what I suggested. My desire is in line with what many other users desire, and what the developers have promised, that with a switch from Xmir to Xwayland among other improvements will return all applications within the Libertine container to full functionality. So going with that route is surely not a waste of development resources, and is within the stated roadmap of development. At that point, there will be both a fully featured Libertine container, and the full suite of UT native apps; perhaps even by then a more fully realised Anbox experience? On the other hand, if you're suggesting that Libertine should just outright be abandoned, well the apps are still exposed in the new OTA-12 drawer and (only some) still launch, which suggests that the developers are not looking at just abandoning this useful application container after users have enjoyed using it for several years.

@TotalSonic said :

UT native apps running on Xenial will give you way better of a tablet experience than Libertine on Vivid ever will.

This seems like an apples and oranges comparison. I think there's value to be had from being able to use spreadsheets on the go, or to design a presentation, or to write up a nice document with a word processor (show me any UT native app for word processing which comes close to the functionality of Libre Office Writer). Clearly UBports would agree, given the pitch for the M10 FHD as "This tablet provides you a great workflow as a portable laptop. Imagine travelling by plane and working on your documents using LibreOffice.". and also it's "The tablet to travel with". That's still up there on the link I cited above. I also recall an independent article strongly advocating for the compact and lightweight form factor of the tablet in laptop configuration, over a bulkier, heavier, noisier, more expensive and less portable in terms of power delivery actual laptop. This was emphasised in the ease of transporting the tablet via carry-on luggage. If I ever find the article again, I will link to cite it. Another fun use case for Libertine is audio editing with programs such as Audacity. Now, this isn't a pure "tablet experience" paradigm, but feels to me like a hybrid, offering portable desktop apps and also great UT native apps (when Libertine gets the dev love it deserves, that is, that's when we expect both of the aforementioned to have equal priority). That in my view would be one of several criteria for the OS to be truly convergent.

Anyway, I recommended that users would be catered for the UT native apps by having an up-to-date, Xenial phone, and this could leave space for a tablet which can run Libertine. It would indeed be a bit of a waste to only have a tablet, just for Libertine under Vivid and missing up-to-date native apps.

@TotalSonic said :

If running desktop Linux apps on a tablet is the priority then the very best way to do it is by using a standard Linux distro on an x86 tablet - I have done this a lot myself - generally running standard Ubuntu with Gnome 3 DE, or by using Xubuntu with the Onboard osk, and Touchegg multi-touch gesture configuration app - and I've set this up on a Dell 7140, Dell 7275, and Samsung Slate 7. The experience on the Dell 7140 has been best out of all these.

For one thing, I can't find any availability for the three tablets you cite on Amazon. For another, I search "(tablet name) price" and I see they vary from £599-999. This is far in excess of what many users would be willing to pay to have desktop apps on a tablet; at that range most would probably reconsider to go for a mid to high-end laptop. By contrast, I parted with £140 including shipping for my most recent M10 HD tablet (which included a superior storage of 32GB, compared to the original M10 HD Ubuntu Touch 16GB edition). A protective case with keyboard and then a mouse on top of that would probably only go to £60, so you're looking at £200 all-in. So much more appealing of a price, wouldn't you agree? For a very reasonable desktop apps experience (under Vivid Libertine, for the time being) which isn't losing out much on that provided by these high-end X86 tablets (and I can state that because I've seen videos showing how the X86 tablets do desktop apps). The Nexus 7 and the Pinetab (leaving aside the fact that the latter will likely never have a Vivid build, for understandable reasons) are also very economically priced, reinforcing the great niche which a Libertine container has for a portable desktop apps experience.

@doniks said :

@TartanSpartan I think it might be worthwile to have Videos of this difference in UX. I mean the X+1 menu story. A video of the device (not a screencapture) might make it clear for (new) developers what was once possible vs today. I never had a M10 and I can't really picture in my mind what you are describing. I understand that its hard to translate menus into a touch interface without writing a new UI. It might help to have that documented

I could do this soon, easily enough, yep. For the reasons you suggest. Albeit I do think many of the developers do understand what the issues are, given their comments about them on Git. But users would also benefit from a clear illustration of the regressions.

I would also caution people to not use the internet where possible for the container, and only with care for when it is unavoidable. The main use case I see for Firefox in such a scenario would be to directly download files to an SD card, for example for locked, premium podcasts; and that's nullified if people are content with using an X86 machine to do this and load the podcasts onto the tablet. It's entirely possible to use VLC, LibreOffice and other useful programs offline. That should remove the ambivalence, I think. In an ideal world I wouldn't have to advocate for this, but Libertine seems to have been de-prioritised ever since Xenial was released for Ubuntu Touch, so again it comes down to user priorities (and this is aided if they run multiple devices for different use cases.).

That's true, I had forgotten about the adaptability of the PineTab for other operating systems, not being reliant on an Android subsystem as the M10 pair and the Nexus 7 is.

Also, it's kind of sad to see a downvote for the initial post. If you disagree so strongly, whoever you are, wouldn't it be more constructive to verbalise why?

TL;DR takeaway: for now, keep your tablets on Vivid for the best Libertine experience (until it becomes usable on more recent versions of Ubuntu Touch), but keep your phones updated to Xenial and primarily use them for mobile apps.

I will preface this post with a series of points which I would argue are very difficult, if impossible, to invalidate. Therefore I hope they can be accepted as consensus, but I am prepared to debate if others have disagreement with any of them, backed up with evidence.

1. Libertine has an important place within Ubuntu Touch and more widely within portable computing. It permits the use of desktop applications on mobile devices which broaden their utility beyond merely mobile apps. Some would argue that a laptop can more capably run any of the desktop apps and even more besides (where Ubuntu Touch may be limited by it's ARM architecture, or lack of packaging such as snaps, or the confinement of the Libertine container, for example), and this is true. However, Ubuntu Touch nevertheless carries the advantages of solid battery life and the ease of charge. Many laptops cannot run for long away from their chargers and a wall socket, even with settings tuned way down. But a Ubuntu Touch tablet has a hefty battery of e.g. 7280 mAh, no fan or hard disk, and efficient power management due to it's ARM CPU. Furthermore, USB slots are becoming increasingly ubiquitous in public transport and places of business (and these can be relied upon without privacy concerns for charging, if the user sheaths their cable in a "USB condom". Ahem... but seriously, that works.), and a USB cable is much smaller in form than a laptop charger. The tablet themselves are also much more compact and lightweight. And a portable USB battery brick or even solar recharging solutions can be used to charge the device for example at a park bench; therefore one can use a tablet for such applications for longer and in more diverse environments than a laptop.

2. Phones are not the optimal use case for Libertine. Yes, users can bring a portable mouse and keyboard, but these may be varyingly clunky, not intended to be carried within pockets like the phone itself, and very few of these have a form factor which fits in flush with the phone and a protective case for it. Even if a user accepts these caveats, they will be left with a screen size which is much too small for a desktop app, unless the phone permits wireless screencasting to a monitor (or, more rarely I think, a wired connection). This would preclude truly on the go use of Libertine, and only a subset of Ubuntu Touch devices do permit screencasting; the MX4 phone for example is excluded.

3. Conversely, tablets are the optimal use case for Libertine. Their screens are sized right, not too far relatively from a laptop screen. Their protective cases often come with a Bluetooth keyboard, and these cases may have stands to prop them up (or else the user can prop them against a vertical surface), approximating the form factor of a laptop whilst adding a touchscreen. The size and touch function together of the screen mean a user may even decide to forgo a mouse, to the extent that the applications they want to use allow an index finger to approximate mouse clicks (and this is an imperfect approximation, but it certainly feels more intuitive than trying to do the same on a phone screen). And as mentioned above, they have strong battery capacities, which always tend to be superior to those of phones. They can even be used with external monitors if this is required, either in a wireless or wired fashion. The strong case for tablets as Libertine devices is reflected in part of UBports' official description of the M10 FHD, "This tablet provides you a great workflow as a portable laptop. Imagine travelling by plane and working on your documents using LibreOffice.". This description is also provided for the M10 HD. I can only presume that the experiences observed on the PineTab and Nexus 7 are similar, or will be at some future date. This is in effect part of a sales pitch by UBports, advocating for the user to acquire a tablet in part because of the superior ability to run desktop apps on them (I'm not saying that UBports will profit from the sale of the devices of course, but you get my essential point here). Note also that the M10 FHD has "promoted device" status, reflecting how much UBports believes in the device as one to highlight for the community.

4. Libertine unfortunately regressed in Ubuntu Touch's jump between Vivid and Xenial, for various reasons. VLC is a strong piece of evidence for this. Before with Vivid, it's menus would layer on top of each other in such a fashion such that Xth menu view would totally obscure X-1th menu's view... which is what you want, because each new menu screen is fully viewable and interactive. Also, you can return to X-1th menu's view by simply closing the Xth, very intuitively. This is completely lost in Xenial, where X will partially obscure X-1, and in so doing X itself will be partially obscured offscreen, with no ability for the user to drag it back into view. This also renders the menu screens partially non-interactive. In short, these windows are not "free-floating". Users can overcome this setback by forgoing the GUI and using CLI config settings, but does everyone want to have to do that? The answer is definitively "no". Such behaviour is also evident in many other programs. Now, let's turn to LibreOffice. All of it's constituent programs "just work" on Vivid, being easily launchable and as responsive as one observes them to be on a desktop or laptop. But not so for Xenial. Logs reveal segmentation faults which prevent some of the constituent programs from launching. Some do launch, and may even allow a sibling to be launched (e.g. launch Writer, then go to File -> New -> Spreadsheet to launch Calc in a roundabout way... maybe, if you're lucky). But different users on different devices have different experiences, with the one constant being that on Xenial, it doesn't "just work" as it did on Vivid. Similar segfaults are observed in other programs, too. Therefore, sadly the sales pitch mentioned in point 3 has been invalidated by the user experience in Xenial, and it will remain invalid unless either users choose not to update or a future update past Xenial corrects for these problems.

With all of those points put together, I would therefore have to advocate that a user considering purchasing a tablet should consciously and deliberately install Vivid for the Libertine experience (and I caveat that this might not even be possible for the PineTab, having likely been developed with Xenial in mind from the outset, and in any event it is currently designated as an experimental device). This would be strengthened if they already have a phone (or a Raspberry Pi, I suppose? Need to think about where this fits into the paradigm) on Xenial, which won't be used for Libertine generally speaking anyway, but more for the mobile apps due to point 2; in this case they would not need an up-to-date experience on the tablet if they consider the desktop apps to be more important for it. They would also then have a small number of pre-installed core mobile apps for the tablet as of Vivid OTA-3, but no OpenStore access, and the legacy browser which in some respects itself saw regressions in the form of Morph browser (but that's a tangent and more relevant for a different line of discussion). They would nevertheless have to be conscious of the fact that the legacy browser is a less security-oriented experience than Morph. Also, the Vivid repositories sadly would not have anywhere near recently updated packages as can be observed in more recent versions of Ubuntu/other Linux distros, but this is another consequence which would have to be accepted. Whether or not the UBports development team and foundation agree with me on this line of advocacy, I hope that the community recognises that this is the most sensible thing for the Libertine experience. I have done this with two M10 tablets (in tandem with more up to date phones) and I have never had reason to doubt my choice in this regard. And I also hope that, recognising these deficiencies observed in Xenial, that the team can target these as they move us toward Focal aka 20.04. It's my understanding that the main thing holding Libertine back is having to move from Xmir to Xwayland. Once that and some other fixes are done, we will hopefully no longer have to use a version of Ubuntu which has been abandoned for over 4 years at the time of writing just to have a usable experience, and have solid and functional desktop app use in a stable release of Ubuntu Touch.

This is great information and thank you guys for the guide. It truly helped me migrate from an M10 with a shattered screen (but the touchscreen still works) to a new one with bigger storage. But I'd like to make my own recommendation on top of this. @thepeter recommends to exclude *data/user-data/phablet/.cache*. This makes sense in general (most app cache directories are pretty disposable and will be rebuilt automatically and as necessary on your new device), but I would take an extra (as in, separate from the main archives you compress) .tgz archive of .cache/libertine-container/ (if you have one or even more than one container) as this will copy over all Libertine apps, app icons, dependencies, the container's rootfs etc. It's arguably just as important as user-data.tgz as that one copies the main Ubuntu Touch apps.

Are there any other directories within .cache which people think would be worthy of backing up and migrating, for genuinely useful and essential stuff like the Libertine apps, and so which should be made an exception to the exclusion command criteria?

@TartanSpartan said in Podbird reignited:

I believe the progression from the Vivid to the Xenial build introduced a regression wherein you used to be able to have your place in an episode automatically saved when the app closed, but now no longer. Shall test this theory on my new M10 (with Vivid) later.

Definitively confirmed.

I believe the progression from the Vivid to the Xenial build introduced a regression wherein you used to be able to have your place in an episode automatically saved when the app closed, but now no longer. Shall test this theory on my new M10 (with Vivid) later.

IMO still the 64GB Meizu Pro 5.

I've decided that I need a small amount of Google on my Meizu Pro 5. I require Revolut and Monzo, two fintech banking apps which pretty much depend on Google services, and I'd much rather run them on that phone with the wonderful UT OS via Anbox rather than an underpowered little Android phone. It seems that both of these apps will require the Google Play Store to fully function. I came across this script:

https://github.com/geeks-r-us/anbox-playstore-installer/blob/master/ubuntutouchanbox-installplaystore.sh

But by comparing it to the x86 equivalent:

https://github.com/geeks-r-us/anbox-playstore-installer/blob/master/install-playstore.sh

I see some differences, e.g. wget is referenced in both scripts but only actually used in the x86 version. Perhaps the UT/arm version is less mature as shown by it's last commit being about 12 months ago on Github. Now, on installing the Google services environment as set up by the script on my Pro 5, the phone no longer runs Android apps correctly (they crash after a few seconds, probably due to memory issues) which suggests that the environment is not correctly installed. Conversely, I have this running fine on one of my x86 installations, and that script seems adequate to such a task.

It is to my understanding that some members of the community have been able to install Google Play Store just fine on their UT devices. Can any of these individuals please provide guidance? Did you have to tweak the installer script, or what? I also note that the libhoudini stuff offered in the x86 script is a red herring for UT, as you of course don't require an arm compatibility layer on arm devices (and while Monzo works on x86 Anbox, Revolut certainly does not). Any help is most appreciated.

I'm glad to come in here now with my experience. Which is exactly the same as described by @ubuntoutou and @matteo ; and specifically I have a MEizu Pro 5 TD-LTE edition. Albeit, as it is my stable daily driver device, I do not choose to flash it to the devel channel and install Anbox in the hope of getting a fringe benefit that it might somehow fix this issue. I have raised this before on the Telegram supergroup, but unfortunately to little avail. I suggested that the donors could crowdfund the developers to obtain all Pro 5 variants, including the TD-LTE version which has perennial availability on eBay, to triage problems which are observed on device flashes apart from the vanilla 32GB bespoke 32GB version. @dobey said to me in response (partially with regard to hotspot problems on the device, but also to this issue): "and let's say some dev agreed to help with the purchase of said device. what happens if they can't reproduce the same issue?. Well, this concern seems nullified given the confirmations of the problem by other users in this topic and elsewhere. Now if he/the other developers want other users to back up the hotspot problems to further justify the purchase of this specific device, well I'm sure those hotspot problems have been cited by other users in this forum, on the supergroup and certainly on issue reporting pages e.g. Github. But I would be happy to help move that forward by summarising the history of those bugs and clarifying confusion/issue drift where ever possible. So most certainly it seems that the next step is to create a funding tier to permit the 64 GB variants of the hardware to be purchased by the devs, and so that they can then look into these problems and hopefully triage them.

How would you rank the priority of bringing Morph browser up to par with what was the standard set by the legacy browser, for OTA-6? Some users cannot upgrade without having a stable and reliable workhorse browser. Will you take requests from them to focus on triage of specific issues for Morph in the release, or for future releases?

@joe I agree. I would like to draw everyone's attention to this issue, and as I say, we need a way to start from fresh with Anbox without having to reinstall Ubuntu Touch. This is especially important for when Anbox is entirely stable and becomes a part of the system image for a future UT OTA.

https://github.com/ubports/anbox/issues/24

Sister issue regarding the importance of being warned or even blocked from installing problematic apps, as desktop snap-based Anbox implementations already do:

https://github.com/ubports/anbox/issues/25

I'm also wondering if Anbox can be destabilised if users try to install APKs beyond a certain storage level? Like, if we install several apps which are each ~100MB, and this is from a starting point of say 1.5GB free storage space left, is there a point at which things start to break? And I mean before running to to 0MB left.

Sorry, but this will put me off upgrading my daily driver to OTA-5 until I have entirely worked through my open history of tabs and exported links/fully sponged out the value of information they have to offer. Unless someone can come up with a way to import those tabs into Morph browser. In the meantime, hopefully a very visible warning (or several warnings across several platforms) can be provided for unaware users.