this is an updated patch attached for 1.7.6~rc1-1.
as i don't intend to send an updated patch for each and every revision,
the current state until it's merged is always available on:
git://vcs.progress-linux.org/packages/git.git
the stuff you want is on the branches progress-artax-backports (f
On 10/20/2010 10:35 PM, Jonathan Nieder wrote:
> install git
> install git-daemon-sysv
> remove git-daemon-sysv
>
> Then git will be installed but git-daemon-sysv will be in the conffiles
> state. Although the daemon is present it should not be run imho.
fixed in updated patch.
Daniel Baumann wrote:
> which package, g-d-sysv (which does check for the existence of the
> daemon), or g-d-run? for the latter, no idea, have no clue about runit.
The former. Try this:
install git
install git-daemon-sysv
remove git-daemon-sysv
Then git will be install
On 10/20/2010 10:20 PM, Jonathan Nieder wrote:
> Should this user be removed in prerm? (Yes, I know git-daemon-run
> does not do that.)
imho yes (which would be a second patch then).
> A general question: what happens when this package is removed but
> not purged? Usually in init scripts the "t
Hi again,
Daniel Baumann wrote:
> newly updated patch attached
Thanks! It's looking better now.
> --- /dev/null
> +++ b/debian/git-daemon-sysv.postinst
> @@ -0,0 +1,19 @@
> +#!/bin/sh
> +set -e
> +
> +test "$1" = 'configure' || exit 0
> +
> +getent passwd gitdaemon >/dev/null || \
> + adduser
On 10/19/2010 07:17 PM, Jonathan Nieder wrote:
> Thanks for explaining; makes sense. It looks to me like the -sysv
> package does not need a gitlog user at all (unlike -run, it writes its
> logs to syslog).
newly updated patch attached (which also removes the replaces:
g-d-r/g-d-s and the depends
Gerrit Pape wrote:
> I don't know about the -sysv package, but for the git-daemon-run package
> this is not a good idea. On deconfigure the saved logs aren't removed
> and still owned by gitlog. On purge the logs are removed and then the
> gitlog user. Removing the user prior to the logs isn't
On 2010-10-19 00:25 +0200, Daniel Baumann wrote:
> g-d-run and g-d- sysv are truly conflicting since both are shipping
> /etc/init.d/git-daemon
No, g-d-run does not ship this file.
> therefore, to my understanding, a conflicts/replaces is needed on both.
The Conflicts is necessary because you d
On 2010-10-19 00:31 +0200, Daniel Baumann wrote:
> On 10/18/2010 10:49 PM, Sven Joachim wrote:
>> I don't think it is necessary or desirable to depend on sysvinit. The
>> init script might just as well be run by upstart or systemd.
>
> then it should be sysvinit | upstart | systemd, but it imho h
On Mon, Oct 18, 2010 at 03:22:36PM -0500, Jonathan Nieder wrote:
> Daniel Baumann wrote:
> > +++ b/debian/git-daemon-sysv.postrm
> > @@ -0,0 +1,10 @@
> > +#!/bin/sh
> > +set -e
> > +
> > +test "$1" = 'purge' || exit 0
> > +
> > +update-rc.d git-daemon remove >/dev/null
> > +
> > +getent passwd gitl
after some upgrade tests: argh, not using debhelper means, that one has
to declare conffiles in /etc manually :/
why are you not using debhelper in the first place anyway?
however, updated patch attached.
--
Address:Daniel Baumann, Burgunderstrasse 3, CH-4562 Biberist
Email: da
here's the updated patch.
--
Address:Daniel Baumann, Burgunderstrasse 3, CH-4562 Biberist
Email: daniel.baum...@panthera-systems.net
Internet: http://people.panthera-systems.net/~daniel-baumann/
>From f83d3746522596ec58fce9dc12f407c2f503012d Mon Sep 17 00:00:00 2001
From: D
On 10/18/2010 10:49 PM, Sven Joachim wrote:
> I don't think it is necessary or desirable to depend on sysvinit. The
> init script might just as well be run by upstart or systemd.
then it should be sysvinit | upstart | systemd, but it imho have a
depends because you can actually install upstart an
On 10/18/2010 10:22 PM, Jonathan Nieder wrote:
> Thanks! I assume you'd be willing to make sure it stays in good
> shape?
i can send patches in case there are updates needed, yes.
>> +++ b/debian/control
> [...]
>> +Conflicts: git-daemon-sysv
>> +Replaces: git-daemon-sysv
>
> Probably a silly q
On 2010-10-18 21:48 +0200, Daniel Baumann wrote:
> +Package: git-daemon-sysv
> +Architecture: all
> +Depends: git (>> ${source:Upstream-Version}), git (<<
> ${source:Upstream-Version}-.), sysvinit (>= 2.88dsf-12), adduser
I don't think it is necessary or desirable to depend on sysvinit. The
ini
Daniel Baumann wrote:
> here's a patch against 1:1.7.2.3-2 which i've applied on my local git
> packages and which i'm using successfully since some time on my own servers.
Thanks! I assume you'd be willing to make sure it stays in good
shape?
> --- a/debian/control
> +++ b/debian/control
[...]
tag 535608 patch
thanks
Hi,
here's a patch against 1:1.7.2.3-2 which i've applied on my local git
packages and which i'm using successfully since some time on my own servers.
I've tried to do it as less intrusive as possible to integrate it with
your packaging style. Please let me know if there'
On Wed, 16 Sep 2009 11:12:40 +
Gerrit Pape wrote:
> On Fri, Jul 03, 2009 at 06:25:21PM +0200, Antonio Ospite wrote:
> > Hi,
> >
> > what do you think about changing the git-daemon-run package to use a
> > standard
> > init.d script? Many of us don't like to depend on runit only for one
> >
On Fri, Jul 03, 2009 at 06:25:21PM +0200, Antonio Ospite wrote:
> Hi,
>
> what do you think about changing the git-daemon-run package to use a standard
> init.d script? Many of us don't like to depend on runit only for one package.
Hi Antonio, please see
http://bugs.debian.org/422139
Regards, G
Package: git-daemon-run
Severity: wishlist
Tags: patch
Hi,
what do you think about changing the git-daemon-run package to use a standard
init.d script? Many of us don't like to depend on runit only for one package.
Maybe this package can be kept and renamed in git-daemon-runit and the default
i
20 matches
Mail list logo
