On Tue, Feb 03, 2009 at 08:01:29PM +0000, Ian Jackson wrote:
> Gerrit Pape writes ("Re: Bug#422139: Please supply a sysvinit script"):
> > A separate binary package named git-daemon-sysv or so, that conflicts
> > and provides git-daemon-run, is the way I'd do the integration.
>
> Personally I think the profusion of tiny packages is not all that
> nice. It would be better to solve that problem (that the
> administrator should be able to choose whether to run the daemon) with
> a debconf question.
When I started to develop an init replacement in 2001, integrating it
into Debian was quite hard, and still is. sysvinit support is hardcoded
into the packages that provide the programs. You can't get the program
without the service enabled through sysvinit automatically. Separating
services from programs on the package level makes different
implementations of service startup and control by different developers
possible, and so helps init replacements. I don't think debconf
question work out that fine.
Furthermore the separation has this advantage
http://lists.debian.org/debian-devel/2005/08/msg01314.html
Yes, the profusion of tiny packages might not be ideal, but in my
opinion it's a reasonable cost for the added flexibility. After all,
the number of packages will increase anyway, and sooner or later we
need to optimize the package database performance.
Regards, Gerrit.
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
