Re: Proposal to add patches to netcfg (#682737)
On Fri, Oct 26, 2012 at 12:56 PM, Philipp Kern <firstname.lastname@example.org> wrote:
> On Mon, Sep 17, 2012 at 11:30:01AM +0300, Sorina - Gabriela Sandu wrote:
>> On Mon, Sep 17, 2012 at 8:26 AM, Christian PERRIER <email@example.com> wrote:
>> > Quoting Sorina - Gabriela Sandu (firstname.lastname@example.org):
>> >>  http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=682737
>> >>  http://anonscm.debian.org/gitweb/?p=d-i/netcfg.git;a=shortlog;h=refs/heads/people/sorina/write_config
>> >>  http://swarm.cs.pub.ro/~sorina/debian/images/08_14_write_config.iso
>> >>  http://projects.gnome.org/NetworkManager/developers/settings-spec-08.html
>> > Just as Phil Kern did for IPv6 support, could you move these changes
>> > to feature/write_config? That would allow me to review debconf
>> > templates without doing so in your "private" branch.
>> Sure, here it is .
>>  http://anonscm.debian.org/gitweb/?p=d-i/netcfg.git;a=shortlog;h=refs/heads/feature/write_config
> it seems that the refactoring for IPv6 broke this patch set in many ways.
> Sorina, do you have time to do a replay of your patches against the current
> netcfg master? If not, please speak up. If so, too. ;-)
Yes, I can do that starting from Thursday, I'll try to bring it to a
mergeable state and also [hope I will be able to :)] add config for
> Currently installation with network-manager is broken because of an additional
> ifupdown option set in /etc/network/interfaces. Given that I introduced that I
> get to keep the pieces, but obviously we should "just" merge your branch so
> that network-manager can get rid of the naïve blacklisting script.
> (Merging seems to be horrible. I think the most sane way is to replay the
> patches through rebasing and then check for errors. The main change is that
> most of the state got encapsulated into a "struct netcfg_interface".)
The fact that the patch relied on global variables didn't look as a
good ideea anyway, so I think changing most of the nm_get* part would
be a big improvement :)
Other than that,I think that the nm-write* part and the logic for the
finish-install script should remain the same. For merging, I would go
with adding the changes manually, somehow seems easier at the moment