On Thu, Jan 15, 2009 at 3:39 PM, Pierre Habouzit <madco...@debian.org> wrote: > On Thu, Jan 15, 2009 at 02:26:18PM +0000, roucaries bastien wrote: >> On Thu, Jan 15, 2009 at 2:44 PM, Pierre Habouzit <madco...@debian.org> wrote: >> > # and to be frank I believe this bug is just plain invalid >> > severity 511687 normal >> > thanks >> >> No the bug is not really invalid it shoke insserver because >> git-daemon-run is a binary file, it does not crash but report loundly >> that it can not read the file git-daemon-run. >> >> Ok it is not a bug per se, but admin could personnalize init.d script.
Yes he could but during install it will shoke innserv. And I dislike to shoke inserv. using a script will also allow to set init priority of this stuff, but i do not really know if runit use it. > So ? The fact that it is a symlink doesn't prevent you from changing it > to a script. Nothing will send a patch i suppose But somebody beat me :) >> > No it doesn't _need_ to, the very standard way to use git-daemon is >> > usually through a super-server. git-daemon-run is just a way to enable >> > git-daemon into runit, which is the packager choice and has nothing to >> > do with the policy as-is. >> >> Yes but put a symlink to a binary file in /etc/init.d is not really >> nice. According to section 9.3, /etc/init.d MUST be script. >> symlink is bad usage at least. > > I'm not sure it must, but what it should and does not, is declaring > /etc/init.d/git-daemon as: > (1) a conffile > (2) not remove it on removal > (3) not overwrite it on install Or use cyril brulebois patch :) >> I believe that this package provide the standalone daemon because it >> wrote to /etc/init.d :( > > You believe or believe*d* ? because afaict git-daemon is part of > git-core. I believed :) I have apt-get --purge remove it :) Will improve cyril patch Regards -- To UNSUBSCRIBE, email to debian-bugs-rc-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
