Purging the F12 orphans
Braden McDaniel
braden at endoframe.com
Fri Jul 24 08:52:48 UTC 2009
On Thu, 2009-07-23 at 23:26 -0700, Toshio Kuratomi wrote:
> On 07/23/2009 06:43 PM, Braden McDaniel wrote:
> > On Tue, 2009-07-21 at 13:35 -0400, Tom "spot" Callaway wrote:
> >> On 07/21/2009 12:06 PM, Braden McDaniel wrote:
> >>> On Mon, 2009-07-20 at 20:11 -0700, Jesse Keating wrote:
> >>>
> >>> [snip]
> >>>
> >>>> Orphan: pcmanx-gtk2
> >>>> gnash-plugin requires /usr/lib/mozilla/plugins
> >>>> gnome-chemistry-utils-mozplugin requires /usr/lib/mozilla/plugins
> >>>> java-1.6.0-openjdk-plugin requires /usr/lib/mozilla/plugins
> >>>> mozilla-opensc-signer requires /usr/lib/mozilla/plugins
> >>>> swfdec-mozilla requires /usr/lib/mozilla/plugins
> >>>
> >>> Umm.... What???
> >>>
> >>> Sigh. pcmanx-gtk2.spec includes:
> >>>
> >>> # We need to own this dir, because we don't want to Requires: firefox
> >>> %{_libdir}/mozilla/plugins/
> >>>
> >>> I think not.
> >>>
> >>> The package already "Requires: xulrunner"; which requires
> >>> mozilla-filesystem, which is what owns %{_libdir}/mozilla/plugins.
> >>
> >> Yeah, that predates mozilla-filesystem. pcmanx-gtk2 is safe to die if no
> >> one wants it (although, having a telnet client plugin for firefox is
> >> somewhat cool).
> >
> > I have no interest in taking ownership of this package; but I would be
> > happy to fix this problem so that the package can be properly culled.
> >
> This can be culled without anything further being done. The
> mozilla-filesystem package provides the same thing as pcmanx-gtk2; when
> pcmanx-gtk2 goes away the dependency will still be satisfied.
Yes, I understand that killing it won't actually break anything.
If it will get removed without fixing this problem, then I will simply
go back to ignoring its existence.
> > Why aren't orphans given open ACLs?
> >
> The argument people have made in the past is that orphaned packages
> should either be unorphaned or be retired; not brought up to snuff for
> random fixes that one developer decides are worthwhile while allowing
> other bugs to be reported without answers, etc.
That's not an argument I necessarily reject; though in this case I was
under the impression that a problem with the package was artificially
prolonging its life. Generally speaking, that sort of thing would be a
problem if no one ever stepped in to fix the breakage. However, if such
packages' lifetimes are limited by other means, then there's no real
problem here.
> (note that almost all packages should be open to provenpackager at this
> point, orphan or not. If you want truly open acls, we need to also
> address the question of how to safely open acls on a package to anyone
> in the packager group.)
I guess I'm just not that cool.
--
Braden McDaniel <braden at endoframe.com>
More information about the fedora-devel-list
mailing list