On Sun, 24 Apr 2011 17:11:10 -0500 William Hubbs <[email protected]> wrote:
> On Sun, Apr 24, 2011 at 04:55:14PM -0500, Canek Peláez Valdés wrote: > > > I think the better way to handle this will be to patch the build > > > systems to not make this an automagic dependency and send those > > > patches upstream. > > > > > > http://www.gentoo.org/proj/en/qa/automagic.xml > > > > > > I'm not a member of qa, but I agree with this position on > > > automagic dependencies. > > > > I'm speaking as a simple user, but I don't think the systemd unit > > files qualify as automagic dependencies as described by the QA > > document. In the first place, as Michael pointed out, we can disable > > them with --without-systemdsystemunitdir, so there is no magic at > > all. > > Ah ok, I guess I missed that. The qa document says there is an > automagic dependency if the builder can't turn this off. But, you are > asying that the builder can turn it off with the --without option. > > In that case, the dependency is not automagic. I'm sorry I didn't explained that clearly. The fact is that it becomes automagic if builder doesn't care about turning it on or off which was the case for many recent fdo package bumps. > > In the second place, the usual Gentoo way of enabling OpenRC > > services is to *add* init.d scripts in the ebuild, and this is > > completely orthogonal to a package installing a systemd unit file > > (the presence of the later does not matter to OpenRC at all). > > This is another topic I may bring up at some point, if you are using > systemd, etc, and openrc is not even installed on your system, do you > need /etc/init.d at all? Once openrc goes stable, I might bring that > up for discussion. I'd say INSTALL_MASK, again. Although it'd be hard to grab all udev rules and other junk like that, it would work fine with most of the scripts. -- Best regards, Michał Górny
signature.asc
Description: PGP signature
