You didn’t understand my comment. I’m also fine staying on Nougat. As long as the patches and fixes get implemented also when LOS stopped doing it for Nougat.
The Fairphone /e/ build is currently on its way to Pie. For a few months we are fine because LOS is on Pie as well. But what will /e/ do when LOS moves to the next version of Android? Similar situations to the FP2 reboot-instead-of-shutdown issue that was never solved on /e/ Nougat are likely to happen.
When Manoj says that /e/ had to fix things inherited from LOS in some occasions, I’m thinking, maybe these fixes relate to versions of Android that LOS was no longer working on, and therefore contained bugs.
I would very much appreciate that. Gael used the analogy of Ubuntu and Debian. Ubuntu does contribute a lot upstream !
Currently, I read seemingly conflicting things along the lines of:
To have your device with /e/ on a newer version of Android, check first that LOS already has that version
Don’t worry if LOS drops support because, /e/ will keep supporting the device.
As far as I know, /e/ for the Samsung S9 is still based on Oreo, an ancient Android version.
What I’d like to see, as LOS is your foundation anyway: contribute upstream to LOS if needed for a specific device, and don’t put the effort downstream for /e/ only. I was very surprised to also see forking e.g. for K9.
And as it is still the beginning of the year, a few suggestions/wishes:
give us a V1.0 for smartphones, and forget other platforms (e.g. laptop) for a while
focus on a limited number of devices and ensure the best experience for those (think Sailfish approach)
put more emphasis on the transparency and the clarity. Like, often, things are stated in the present tense: “/e/ offers this and that”. And when it doesn’t work, we are reminded “don’t forget that it’s still beta”
I was doing some digging around checking devices on the Lineage supported list and /e/'s list
Lineage has support for a large number of smartphone and tabs.
The /e/OS list is small in comparison. Also there are about 43 phones which are in the /e/ list but for which Lineage has been dropped support. Still need to verify all the details.
Got you, although I’d still be relieved that I can reach out to someone here and see if there are any plans on dropping support completely.
That’s good to know, unfortunately that doesn’t apply in my case.
It has been said in other threads but I’ll still give it a +1. I also like the idea of contributing upstream to LOS, isnt’ it more efficient this way for everyone since other developers may be encouraged to pick up ‘older’ phones if /e/ sitll shows some interest in them?
Ancient? I have a old device “Klipad pouces” running Googles Android 4 KitKat! Still works, so why trash it? People are right about LOS, they drop support for older devices too often too soon, my tablet “Samsung Galaxy Tab 2 7” worked on an official Lineage, but for some reason LOS dropped the support after Lollipop, and right now it works on an unofficial build with Android 6 Marshmallow. Works perfectly, so why throw it away?
Sustainability means taking care of your devices, and they are only obsolete when it doesn’t work anymore.
Thanks for compiling this! However, I think there is some confusion on what “supported” actually means. @Manoj Could you maybe shed some light on this?
For me this would mean:
images are “officially” (i.e. on e-Servers) build and provided via OTA, as long as they build without error (this is what the LOS guy calls “blind builds” in his post). In case of error, someone from the /e/ team takes a look and fixes it so it builds again.
Generic security fixes (i.e. not device-specific) are included “automatically” in the build as long as the underlying LOS source is maintained.
/e/ specific fixes that are required by changes on the underlying LOS source are applied (these arealso generic for all devices I’d imagine).
Device specific security fixes are included in the build. BUT: what would that actually be, swapping binary blobs for newer versions from the vendor? I cant think of anything else, but this might be the reason why LOS drops support, as someone needs to isolate, test and prepare this. But then, I can not image the /e/ team handling this for a large number of devices.
So my guess is 1+2+3, with 4 being not covered by /e/ nor by LOS (as unsupported devices don’t have a maintainer). Or do I miss something on (4), like changes in the device tree config that might be needed?
About contributing back to LOS: any fixes that are done at (2) can/should be easy to pass on upstream. Also bugfixes from (3) should mostly be ok to contribute back, unless they are just workarounds to make the e-code work.
As others have pointed out, there are unofficial / unsupported builds for quite old devices that are still usable. As long as they build (i.e. with e-docker-script or LOS-docker-script for that matter) are they “secure” or potentially vulnerable since (4) above is missing?
I guess we won’t reach a common understanding anyway. So, a first feedback. There is a list of “supported” device, but there is only a beta /e/. Basically, as an OS, in my eyes, /e/ doesn’t support any device yet. Not until /e/ reaches 1.0. So, to me, you now have a target device list. Anyway…
As far as I’m concerned, in 2020, support includes security support. Based on https://source.android.com/security/bulletin, Google dropped support for Nougat in August/October 2019 (first 7.0.x, then all 7.x). Is there a list of CVEs that /e/ has fixed since October 2019 for those supposedly supported HTC devices ?
The same applies to the Samung S7 that you sold and still sell pre-installed. It seems to also be on Nougat e-0.7-n-2020011737872-dev-hero2lte.zip. What about its security updates since October 2019 ? I hope the upgrade that is being tested will be proven OK.
I actually appreciate the hard work. I’m not fan of the expectation management which seems unnecessarily enthusiastic to me. For instance, you could focus on support for just a handful of devices; not 80+ devices, including some still on Nougat. For instance, when buying a pre-installed phone, the site says “Tested 100% functional”. I’m not sure it mentions “Beta”.
I don’t know enough about how the Android system fits together, but this point is a concern for me. GrapheneOS which is security focused, dropped the Pixel original as soon a Google dropped support for it. This was due to no longer being able to ensure security. Depending on how you see the privacy threat, there are 2 main attackers, the “trackers” which /e/ is focused on, and the hackers who rely on security bugs. Ideally I want an OS which prevents both types of information gathering. I don’t know how secure a device without the device specific updates remains.
I’s say Google is not the all important factor here, but rather if LOS still backports security fixes to the 14.1 (nougat) branch. They do not do so “officially” (whatever that means), but in the build-changelogs there are constantly community based security-backports for LOS14.1 visible (see https://review.lineageos.org/q/branch:cm-14.1)
However, somebody would have to watch what those fixes cover and if they are “generic” (eg ssl-library updates) or device-specific (eg fixing kernel vulnerabilities) and add what is still missing (based on what, CVEs?) to ensure a fully “supported” device.
@waluwaz: I am with you here concerning the number of devices, if my assumptions on what is needed to keep a device properly supported are true there is no way this could be done by the /e/ team alone for all those devices that where dropped by LOS. Hence the need to collaborate and spread this work over more shoulders.
One more thing that will differentiate /e/ from LOS (hopefully coming this year?) will be the ability to upgrade the /e/ version to the next Android base version with OTA updates. Currently for LOS to upgrade you must reflash the new ROM. This is too much to expect from the /e/ target audience.
Once this feature is working then we will see devices being move to parity with LOS base Android version.
The answer from Gaël Duval was interesting to me because it contains more details than what is available on the official /e/ website.
I myself fully support the idea of making an alternative for the ordinary (casual) users. And I think /e/ is doing it the right way (cloud integration, long-term support).
I agree with some comments here pointing that /e/ should focus on a limited number of devices and ones that will be in mass use for some years to come (and obviously those which are durable by design, notably the Fairphones).
Edit : and devices that /e/ can sell to people (ordinary people don’t flash devices).
The trouble seems to me that groups of Open Source developers are each pursuing their own goals and projects instead of uniting. Therefore each group has too few resources…
By the way, could /e/ hire the people who make “Lineage for microG” since this appears more close to /e/ ?
I used LineageOS for several months last year. My impressions:
As an operating system it was excellent.
The installation instructions for me as a newbie were great.
There seemed to be nearly zero communication between the developers and the users.
It was unclear to me if anyone was planning any future expansion of the project.
/e/ was trying to build an app store and an ecosystem so I gave it a try. I found that as a daily driver it works fine for me. OTA updates come fairly often. I am entirely satisfied.
Did you buy the refurbished phone or flashed it yourself? Were the instructions also easy to follow? Any other issues or things you wish were different in /e/OS?
I bought a second-hand Nexus 5x and flashed the ROM. I followed the instructions carefully and everything went really smoothly. Later I used the same instructions but with the /e/ xxx.zip file. Same good results.
/e/ has a few quirks. The substitute TTS engine isn’t nearly as good as Google’s, but other than that I don’t really see any disadvantages to going Google-free.
