[Replicant] making non free versions of replicant

[Denis 'GNUtoo' Carikli]

On Thu, 14 Nov 2019 01:12:34 +0200
"ni nhar" <ninhar at inbox.lv> wrote:

> And maybe nothing is gained from a repnonfree system compared to a
> lineage os system.
A way to deal with it would be to work with various upstream, which
include LineageOS and GNU/Linux to push our implementations there.

For instance if libsamsung-ipc and libsamsung-ril get used by LineageOS:
- It would make upstream more free
- We would also share maintenance
- People could add support for more devices in it

As for GNU/Linux we already started contributing to the Linux kernel as
part of porting Replicant to Android 9.

We could also try to add support for libsamsung-ipc in ofono as there
was a patch that didn't get merged because of licensing issues that are
now resolved.

Collaborating with GNU/Linux would enable GNU/Linux distributions to
support the devices supported by Replicant as well, and we would be
able to support GNU/Linux smartphones more easily.

It also helps with Replicant development as you have access to many of
the GNU/Linux tools.

For instance, I use Parabola for testing Linux patches on the Galaxy
SIII, so I just need to cross compile the kernel, and rsync the image +
dts + test scripts to the device and reboot it. Then I ssh in it and run
the test scripts.

> Wrongly I had the impression that replicant search for unwanted
> monitoring pieces of software and remove it should they find
> something. Rather if replicant gets aware of a piece of unwanted
> monitoring software, it gets removed.
You could either probably get it removed without having to fork
LineageOS, or just advise users on how to deactivate it.
You can also advise users not to install google play.

There are also pieces that we probably don't know about that are
automatically removed as we don't use nonfree userspace drivers for the
GPS, 3D acceleration, etc.

It would be interesting to have more information on that, for instance I
wonder if there are known cases where the nonfree GPS driver sends back
user position somehow, or if there are no known cases.

For GPS, monitoring may also exists in relation to the supplementary
AGPS data that is sent to the GPS as it's often transited over a
network in the case of nonfree implementations.

In free implementations (like for the Openmoko) that data was saved and
restored to get a fix faster, so it didn't require network access.

> Isn't postmarket os far away from being able to provide usable
> phones? 
You could look at the status on their website. Last time I looked they
managed to boot many devices but if my memory is good, none were able
to do calls and SMS. The status of the individual devices are on the
wiki but I don't know how up to date it is. The good thing is that they
are probably working with upstream as well.

We need to meet the people behind postmarketOS to see if we can improve
the cooperation between Replicant and PostmarketOS. Right now
postmarketOS probably also contribute upstream so we collaborate
indirectly in this way. We also look at each other wiki.

> My understanding is should skilled replicant people decide to
> build repnonfree, then the time frame would be manageable. Is that
> correct? Or we do not know how difficult it will be to make a
> repnonfree version?
I will not participate in making a nonfree version of Replicant, but I
do participate in upstreaming code, and I'd be happy to collaborate
more with other projects to do that if I can manage to find the time
for it.

Denis.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 833 bytes
Desc: OpenPGP digital signature
URL: <http://lists.osuosl.org/pipermail/replicant/attachments/20191114/e68ce51c/attachment.asc>