Anbox brings Android apps to the Librem 5:
Anbox brings Android apps to the Librem 5:
Well, this is fantastic.
Can someone smarter than me explain this point:
- Switch to iptables-legacy
Also, I’m curious why Squeekboard doesn’t work well enough to use that in Anbox.
I love how little fuss this has been delivered with. Shrugging off big tech. “Oh, it’s no big deal. Just Android app compatibility. I doubt many people are interested in that.”
Out of interest, which apps are shown in the video? I am clueless when it comes to Android, so I can’t identify them. That browser looking thing had tabs that look a bit like Firefox tabs, so I guess that’s Firefox?
That is being removed as a requirement (I’ll update the post). It is needed to run ATM, but it won’t be in the near future.
iptables’ successor nftables has become/is becoming the default. iptables-legacy is the name of the package you have to install if you require legacy iptables instead of nftables. (AFAIK)
Here is a list of apps shown in the video: Signal, Privacy Browser, Vector Pinball, Antimine, Anuto TD, Element, F-Droid, Notes.
This is great news with the coming arrival of Evergreen. I have only 1 or 2 android things I want (such as the ABC listen app in Australia) so this just tops things off nicely. Thank you for the effort Purism.
Why a reboot is needed to complete installation process?
It seems crowded on the blog today
I agree with @patch that this could have been made into a bigger deal. But on the other hand, I’d be wary on how much you’d want to associate [auto-correct suggested “assassinate” here ] L5 with Android overall (so, as not to be thought of as one of the main features), and that Android apps should be understood to be used only when safer native alternatives are not available.
This is excellent.
Didn’t really need the lecture on using proprietary apps, but you know, as long as I can do whatever I want on the phone, it’s all good.
Yes, I agree. I think it’s been pitched just right.
It’s also not a fully polished solution at this point.
Only problem is, I will now have to investigate individual Android apps I encounter, to see whether they are acceptable, rather than just having a blanket policy of “Eh, can’t use that. It doesn’t support my phone.”
A comparison would be like the 45 year-old HP3000. Any purchased software outside of the fundamental operating system (like networking or a compiler) wrote to the system library (SL), which is a file. But the modified SL only got read into memory at reboot time. Hence the reboot if you wanted anything useful.
I think the principle is the same if the details of how it is done are different in today’s tech.
(And who would have thought ‘networking’ was an add-on product today ? But it was in the 1970’s)
This is great news as there are a couple pieces of software I require that don’t have Linux equivalents, like Okta Verify. I was just going to keep around my old phone for these but now I might be able to totally ditch that phone.
What I suspect is, the installation scripts tweaks some kernel parameters in /etc/sysctl.conf and friends, which are read only on boot, and also adds some kernel modules (whatever anboxy needs to run in kernel space) somewhere in /etc to also be loaded on boot. Then a reboot is a least effort way to make them active. The script could also set the parameters on a running kernel and load necessary modules (this is what startup scripts do, among other things) and lo, no reboot would be necessary.
But there are often obstacles to this non-reboot approach. So I ask. I hate reboots. I hate when I need to interrupt running applications in order to install something completely unrelated.
I find it somewhat ironic that what was an unmet stretch goal during the crowdfunding campaign ended up being finished before getting the camera working.
I guess this just goes to show how much harder low level stuff is.
As if the existence of every Electron.js application I know of wasn’t a testament to that.
Yeah, a kernel parameter. Same concept as an SL on the old HP3000. It got tweaked, so reboot.
Those were my sentiments.
Yes, it’s great that some ruly truly Android/spiPhone-only apps can now be run from Day 1. But it may discourage people from using and/or creating more secure, more private native alternatives.
I wouldn’t have minded a period of time between completed Evergreen delivery and Anbox delivery where people get to think about and find out what apps they really need.
Not necessarily. Most kernel parameters are changeable at any time.
Yes I imagine nowadays are different. Who would have thought hot swappable components back then?
On the contrary, ignoring the large swathe of apps that are already customized for mobile and hoping that a relatively small crowd of developers that are targeting Linux will continue further and target Linux mobile seems like an even more foolish endeavor.
The Librem needs to go mainstream. It doesn’t do that hoping for things that are very unlikely to ever happen without a paradigm shift to spur it.
Instead, the mini lecture in the video is Purism positioning itself and the phone to appeal to those small group of developers as you are hoping. (I still say that labeling all proprietary software as bad is wrong)
However I think things like the Wine project are probably the most important things going on for Linux in general.
This Anbox on the L5 was not only a good idea, on the part of Purism, but an absolutely essential one.