[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: let's split the systemd binary package

previously on this list Olav Vitters contributed:

> On Tue, Oct 29, 2013 at 06:37:35PM +0000, Kevin Chadwick wrote:
> > Of course they do even if the couple of people possibly concerned with
> > it that I know use.. is it Citrix? I was merely pointing out that it
> > is an extremely small minority of Debian users but possibly? a majority  
> Do you have any references to back up how you know this? Or just merely
> guessing? It seems like pure guesswork.

Until someone points out some new functionality I have missed, this is
surely obvious especially if the *buntus are included.

> > This should be considered in weighting the pros and cons that's all
> > especially when terms like real features (largely gone undefined) are
> > being banded around. As I have said issues that affect many and people
> > may actually notice have gone are easily fixed as far as I am aware
> > (certainly the ones mentioned like suspend, as I do so when disabling
> > polkit very easily without compilation). So how many debian Gnome users
> > will notice the breakage aside from suspend and ? how many will
> > continue to use Gnome if the default is changed as has already been
> > raised.  
> Are you taking up ConsoleKit development or not? Loads of things could
> theoretically maybe be done. What matters is something concrete.

I have no use for consolekit and it doesn't run on my systems and none
of my users notice, so why would I?. What I don't agree with is this
ultimatum that something must be done and making a mountain out of a
trench. Worst case is holding systemd back or using consolekit and all
this may be solved in x number of unknown ways by the time it matters to
stable. I'm also sure those that need session tracking could easily
afford to sponsor the work if they need it and use Debian.

> > On top of that, large organisation's should have no problem solving
> > this and do they use debian or want support from Red Hat/Citrix in
> > most cases?  
> Please don't turn this into a Canonical vs Red Hat thread.

I am not, I have spotted and cited trends and made statements perhaps
without citing other annoying trends in detail that may alter my tone
to only parts of RedHat (such as documentation, configuration
methods...). I respect the company as a whole and don't forget many
Redhat employees have publicly spoken out against systemd. I was merely
responding to the post of lennart's that may make many think there is no
alternative, when they specifically have been talked about recently.
There are always alternatives. Who knows even linux itself may fork one
day but I hope not.

> > I don't need the dbus system bus personally either but I understand the
> > vast vast majority do in current setups, so that is a real issue of the
> > future if permitted to land into the kernel as the only option (I doubt
> > it) and as Canonical/Ubuntu and Google have concerns on multiple fronts
> > here I think it is certainly worth waiting that out and should not
> > really be used as an argument currently.  
> GNOME relies on d-bus for various things. Not liking d-bus doesn't
> change that fact.

Who said I didn't like dbus. I even said the session bus should be used
where it is best suited. I do think it is often used when it shouldn't
be and do disagree with parts of dbus. dbus has atleast 3 major distinct
functions that I can think of. Running programs as root is one I
completely disagree with and with evidential good reason.


'Write programs that do one thing and do it well. Write programs to work
together. Write programs to handle text streams, because that is a
universal interface'

(Doug McIlroy)

In Other Words - Don't design like polkit or systemd

Reply to: