Nylon liveth (but on in OE?)

Nylon liveth (but on in OE?)

[Martin Dietze]

Nylon is the distribution with which the 4G Meshcube/AccessCube
was shipped. As a former 4G employee I took over maintainership
and started re-integrating Nylon into org.openembedded.dev in
2009. However Nylon was and still is based on an old toolchain
and a 2.4 kernel. This lead to increasing problems keeping the
code buildable, as more and more code would no longer work on
a basis that old. Thus Nylon is now obsoleted in OE (and I'd 
say rightly so).

Not having had the time to work on this in the last years I've
now decided to go the easy way and simply fork at some time in
2009 when the code was still more or less working and continue
from there. The main benefit from using my work rather than the
original 4G toolchain (which was itself a fork from a much older
OE snapshot) I see in having a larger set of recipes to add
components from if needed and also some updated components in
the base system

This work has now been done, and I wonder where the code should
live from now on. I would thus like to ask if keeping my branch
in the OE repo would be an acceptable option to the project
maintainers. Doing this we could probably just remove all the
remaining mtx-[12] artifacts from the .dev branch (I am not sure
if mtx-3 is maintained by anyone any longer, at least it is not
by me).

I know that there's possibly only very few people interested in
this, but I'd really like to avoid forking completely, also 
the idea of still being able to grab some recipes from the .dev
branch if needed would be nice, and it would simplify contributing
for others. If this idea is not an acceptable option for you, of
course, I can still push my fork to github.

What do you think about it?

M'bert

-- 
----------- / http://herbert.the-little-red-haired-girl.org / -------------
=+= 
Wie verhindert man, dass einem die Geige gestohlen wird?
Man legt sie in einen Bratschenkasten...

Re: Nylon liveth (but on in OE?)

[Paul Eggleton]

On Saturday 02 March 2013 18:46:24 Martin Dietze wrote:
> Nylon is the distribution with which the 4G Meshcube/AccessCube
> was shipped. As a former 4G employee I took over maintainership
> and started re-integrating Nylon into org.openembedded.dev in
> 2009. However Nylon was and still is based on an old toolchain
> and a 2.4 kernel. This lead to increasing problems keeping the
> code buildable, as more and more code would no longer work on
> a basis that old. Thus Nylon is now obsoleted in OE (and I'd
> say rightly so).
> 
> Not having had the time to work on this in the last years I've
> now decided to go the easy way and simply fork at some time in
> 2009 when the code was still more or less working and continue
> from there. The main benefit from using my work rather than the
> original 4G toolchain (which was itself a fork from a much older
> OE snapshot) I see in having a larger set of recipes to add
> components from if needed and also some updated components in
> the base system
> 
> This work has now been done, and I wonder where the code should
> live from now on. I would thus like to ask if keeping my branch
> in the OE repo would be an acceptable option to the project
> maintainers. Doing this we could probably just remove all the
> remaining mtx-[12] artifacts from the .dev branch (I am not sure
> if mtx-3 is maintained by anyone any longer, at least it is not
> by me).
> 
> I know that there's possibly only very few people interested in
> this, but I'd really like to avoid forking completely, also
> the idea of still being able to grab some recipes from the .dev
> branch if needed would be nice, and it would simplify contributing
> for others. If this idea is not an acceptable option for you, of
> course, I can still push my fork to github.
> 
> What do you think about it?

I don't have a personal interest in this particular distro/hardare, but if the 
fork you have made is working then I can't see a compelling reason not to host 
a branch for it in the OE-Classic repository; at least that keeps things more 
or less in one place.

However, is there any chance you'd be able to look at moving up to be based on 
top of OE-Core which is where OE is at currently? Obviously toolchain and 
kernel would be the first hurdles. For bringing any needed metadata up-to-date 
though the community can provide help where needed.

Cheers,
Paul

-- 

Paul Eggleton
Intel Open Source Technology Centre

Re: Nylon liveth (but on in OE?)

[Martin Dietze]

On Tue, March 05, 2013, Paul Eggleton wrote:

> I don't have a personal interest in this particular distro/hardare, but if the 
> fork you have made is working then I can't see a compelling reason not to host 
> a branch for it in the OE-Classic repository; at least that keeps things more 
> or less in one place.

I have for now created a personal branch `mdietze/nylon' for my
changes. 

> However, is there any chance you'd be able to look at moving up to be based on 
> top of OE-Core which is where OE is at currently? Obviously toolchain and 
> kernel would be the first hurdles. For bringing any needed metadata up-to-date 
> though the community can provide help where needed.

The last time I managed to create a buildable and working
Nylon firmware image based on an up-to-date OE.dev was in
April 2009. After that my build broke practically every other
week as my distro's dependencies and particularities were just
too far beyond scope for practically everybody else.

Since do not have enough (spare) time I can invest in this,
forking seemed the only reasonable way. As long as nobody ports
the system based on 2.6 kernel from OpenWRT to Nylon on OE, I
don't really see any benefit from keeping the build alive by
permanently adapting things to recent changes in OE (also 2.4
kernel based distros are obsoleted in OE anyway).  

Things may look different if people actively using Nylon on
AccessCubes or Meshcubes want to join in :)

Cheers,

M'bert

-- 
----------- / http://herbert.the-little-red-haired-girl.org / -------------
=+= 
My name is spelled Luxury Yacht but it's pronounced Throatwabbler Mangrove.