[systemd-devel] Have custom agetty behaviour even after upgadres

Colin Guthrie gmane at colin.guthr.ie
Wed May 16 03:56:26 PDT 2012


'Twas brillig, and Dave Reisner at 14/05/12 22:31 did gyre and gimble:
> On Mon, May 14, 2012 at 11:01:53PM +0200, Lennart Poettering wrote:
>> > On Mon, 14.05.12 14:51, Dave Reisner (d at falconindy.com) wrote:
>> > 
>>> > > 
>>> > > On Mon, May 14, 2012 at 08:39:23PM +0200, Lennart Poettering wrote:
>>>> > > > On Thu, 10.05.12 23:27, Seblu (seblu at seblu.net) wrote:
>>>> > > > 
>>>>> > > > > Hello,
>>>>> > > > > 
>>>>> > > > > on my archlinux test computer i would have first console not cleaned
>>>>> > > > > and other spawned statically (crazy idea isn't it).
>>>>> > > > > 
>>>>> > > > > So I've turned NAutoVTs to 1 in /etc/systemd/systemd-logind.conf.
>>>>> > > > > I've copied /usr/lib/systemd/system/getty at .service into
>>>>> > > > > /etc/systemd/system/ and patched as following:
>>>> > > > 
>>>>> > > > > My issue is at every systemd package upgrade, getty at tty1.service is
>>>>> > > > > replaced by a new one linked to
>>>>> > > > > /usr/lib/systemd/system/getty at .service.
>>>>> > > > > And i loose my configuration.
>>>> > > > 
>>>> > > > This sounds like a packaging problem downstream, nothing we can fix upstream.
>>>> > > > 
>>> > > 
>>> > > Strange that you say that, since make install does the following:
>>> > > 
>>> > >   ( cd $(DESTDIR)$(pkgsysconfdir)/system/getty.target.wants && \
>>> > >     rm -f getty at tty1.service && \
>>> > >     $(LN_S) $(systemunitdir)/getty at .service getty at tty1.service )
>>> > > 
>>> > > See line 3128ish on git master. This is absolutely something that we (Arch
>>> > > Linux) inherit from upstream.
>> > 
>> > "make install" is not really too useful for end-user installs, is it?
>> > The distro package manager should handle configuration files specially,
>> > the way .rpm or .deb do it.
> I'm saying 'make install' from within the packaging environment with
> some sort of DESTDIR= variable pointing to the packaging root. Fedora
> does this, Mageia does this, Arch does this. Anyone who tries to package
> something this complex on their own is doing it wrong.

So are you suggesting that if DESTDIR is set, the /etc/ symlinks should
be skipped?

I'm not sure I'd vote for DESTDIR dependant stuff - that's just a little
bit too cryptic and subtle when working on code for packaging vs working
on code for hacking.

Personally I'm quite happy with the simple packaging tweak to clean them up:
http://svnweb.mageia.org/packages/cauldron/systemd/current/SPECS/systemd.spec?revision=235024&view=markup#l256

(s/used/users/ in the comment)


Col




-- 

Colin Guthrie
gmane(at)colin.guthr.ie
http://colin.guthr.ie/

Day Job:
  Tribalogic Limited http://www.tribalogic.net/
Open Source:
  Mageia Contributor http://www.mageia.org/
  PulseAudio Hacker http://www.pulseaudio.org/
  Trac Hacker http://trac.edgewall.org/



More information about the systemd-devel mailing list