[Replicant] Anbox

W. Kosior koszko at koszko.org
Sat Jun 19 14:59:54 UTC 2021


Thank you for the responses. I hope the rest of list users didn't have
problems reading my email - it seems osuosl.org breaks DMARC. I'll
contact admins and maybe add spf entries for osuosl.org's agents if
problem persists.

Back to Android.

> This may be somewhat off-topic for the Replicant list, but anyways

> As for the first three, you can get rid of them by running pretty much
> any community build of Android (some may still come with GApps due to
> popular demand, so look out for that).

Can I be 100% certain the GApp-less Android _userspace_ from, say,
LineageOS is going to be 100% blob-free? Isn't there any other nonfree
thing that might have slipped through? I thought about using Replicant
and not Lineage purely because of that.

Also, what about anti-features? Is it possible that some components of
the GApp-less userspace, although freely licensed, will still be calling
home to Google? I guess that's relatively easy to work around but I am
asking anyway, just to know.

> For PC hardware, things look a bit better [...]

Btw, I am on RockPro64 (using it as my daily driver now). Things look
even better here (no Intel ME, etc.). With nonfree drivers/firmware
problems gone, I thought it would be relatively easy to get a deblobbed
Android+Anbox for my platform. Now I see it might be a bit difficult.

Also, than you for making me realize that x86-Android and Anbox's
bundled images are outdated. I didn't notice that.

> You would not need an Android image, as Anbox already comes with the
> full Android stack to run on top of your existing Linux installation.

Well, I admit I was thinking I can *just* grab an existing Android
image of respective processor architecture (aarch64 for RockPro) and put
it into Anbox. It wouldn't work, would it? Still, the images provided by
Anbox don't seem as trustworthy as I would like - Google's default build
procedure they seem to have used is "dirty", i.e. uses prebuilt JDK and
Make bundled with sources.

> At that point, I wonder why Replicant is not offering an x86 port.
> It’s a different processor architecture, but most devs would already
> have hardware for testing, and obtaining free hardware drivers should
> be much easier than for most handheld devices...

I agree! Or why not ports for libre SBCs like those from PINE64 and
Olimex? Either way, it seems this could be made relatively easy by
copying some code from projects that already did the exact same thing
with stock Android.

> One of the comments from your same message might as well apply here:
> to offer a x86 port, Replicant and the work group of FSDG-compliant
> distros need to audit Anbox in regards to whether it does follow the
> FSDG. So all it takes is people like you and Wojtek, or someone else,
> to start working on this review with the aforementioned work groups.

Is there a misunderstanding? Anbox and an x86 Replicant port
seem to be different issues. Sure, once we have an x86 port we
could use it to provide a Replicant image for x86 Anbox - but that
doesn't mean Anbox has to be audited or even exist for an x86
Replicant port to be possible.

As to review - I am sorry, I cannot spare time for that. I already have
more daring freesw work to do[1][2].


Wojtek

[1] https://trisquel.info/en/forum/librejs-ranting#comment-158316
[2] https://git.koszko.org/browser-extension/ (sorry for advertisement)

-- 
website: https://koszko.org/koszko.html
PGP: https://koszko.org/key.gpg
fingerprint: E972 7060 E3C5 637C 8A4F  4B42 4BC5 221C 5A79 FD1A
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 228 bytes
Desc: OpenPGP digital signature
URL: <http://lists.osuosl.org/pipermail/replicant/attachments/20210619/bdfc8c48/attachment.asc>


More information about the Replicant mailing list