[libvirt PATCH] meson: Fix libvirtd|virtproxyd socket prefixes

[Erik Skultety]

On Tue, Aug 04, 2020 at 10:00:19AM +0200, Pavel Hrdina wrote:
> On Tue, Aug 04, 2020 at 09:48:58AM +0200, Pavel Hrdina wrote:
> > On Tue, Aug 04, 2020 at 09:30:42AM +0200, Erik Skultety wrote:
> > > For the daemons in question the correct socket prefix is "libvirt-",
> > > not "libvirtd-".
> > > 
> > > Fixes: dd4f2c73ad7f9fc0eae5325d5bf5786afd3a467e
> > > 
> > > Signed-off-by: Erik Skultety <eskultet at redhat.com>
> > > ---
> > >  src/remote/meson.build | 4 ++--
> > >  1 file changed, 2 insertions(+), 2 deletions(-)
> > > 
> > > diff --git a/src/remote/meson.build b/src/remote/meson.build
> > > index 25712c943b..5983238a0a 100644
> > > --- a/src/remote/meson.build
> > > +++ b/src/remote/meson.build
> > > @@ -184,7 +184,7 @@ if conf.has('WITH_REMOTE')
> > >        'service': 'libvirtd',
> > >        'service_in': files('libvirtd.service.in'),
> > >        'name': 'Libvirt',
> > > -      'sockprefix': 'libvirtd',
> > > +      'sockprefix': 'libvirt',
> > >        'sockets': [ 'main', 'ro', 'admin', 'tcp', 'tls' ],
> > >      }
> > >  
> > > @@ -218,7 +218,7 @@ if conf.has('WITH_REMOTE')
> > >        'service': 'virtproxyd',
> > >        'service_in': files('virtproxyd.service.in'),
> > >        'name': 'Libvirt proxy',
> > > -      'sockprefix': 'libvirtd',
> > > +      'sockprefix': 'libvirt',
> > >        'sockets': [ 'main', 'ro', 'admin', 'tcp', 'tls' ],
> > >        'deps': libvirtd_socket_conflicts,
> > 
> > How did you figure out that the prefix is libvirt? It's not correct.
> > Check the autoconf version or even in your fedora:
> > 
> >     rpm -ql libvirt-daemon | grep ".*\.socket"
> > 
> > and you will see that it's libvirtd.
> 
> So I double checked is as it looked strange to me and even though the
> service and the systemd socket files use libvirtd the runtime unix
> socket has libvirt as prefix. Sigh.

Well, historical reasons... we couldn't have renamed it with the modular
daemon introduction, because that would have been a regression.

> 
> So the patch is correct, sorry about the noise.
> 
> Reviewed-by: Pavel Hrdina <phrdina at redhat.com>

Thank you.

Erik