[pkg-fso-maint] GSoC 2011 - project ideas?

[Thibaut Girka]

Le mercredi 16 mars 2011 à 13:28 +0100, Gaudenz Steinlin a écrit :
> Hi
> 
> I was one of the two mentors of Thibaut's last summers project.
> 
> Excerpts from Steffen Möller's message of 2011-03-15 12:55:57 +0100:
> > 
> > > As a quick sum up, with my modified code, d-i can run and install
> > > Debian on the FR. Some of this code is now in d-i, but some parts
> > > (the gta02 target itself, kernel things, and the uboot-env thing)
> > > are not.
> 
> Do you have an overview of what's left to be merged? Is there an easy
> place to download your patches or did you submit everything as bug
> reports? If things can't go upstream (at least into branches) the next
> best thing would be to have everything as patches in bug reports and
> to have a usertag to easily list all these reports.

I've attached most (but not all) patches in bugs on the BTS (under
usertag[1].
Everything else is in the pkg-fso/d-i.git repo.

> > > I've dropped kernel upstreaming for now, but I've started an
> > > on-screen keyboard project so that g-i could be used on the FR or
> > > any other touchscreen-only device.
> 
> Do you have an overview of what's left to upstream to have very basic
> support to install on the network-console?

glamo-core and glamo-mci (+ platform-specific code) should be enough,
but that's already some work.
The device would then run as a headless machine, without wi-fi support,
not sure many users whant that, but it'd be a start.

> > > I've not really know much on the GTA04, but I suspect kernel
> > > upstreaming would be easier as things like the glamo chip have
> > > been removed.
> > >   
> > Could you possibly mentor something? You certainly have a good overview
> > on what would be feasible to aim at. D-i has recently moved towards git,
> > so maybe this could help us in some way.
> 
> I would also be available to mentor something again this year. IMO as
> long as there is now kernel with at least basic support in Debian
> there is not much point in doing more d-i work as it won't get
> integrated into any official release anytime soon. We don't have to
> upstream everything! It's enough to have basic support to bootstrap
> the system and after that people can install a better suited kernel
> from the pkg-fso repository if the want. AFAIK the major part missing
> is the glamo mfd module for SD card access.

I agree, more d-i work for the FR is useless until at least basic
support for the FR is in mainline kernel.