[libvirt] [PATCH v3 15/48] build: don't hardcode /etc in the config related files

[Andrea Bolognani]

On Mon, 2019-07-29 at 18:10 +0100, Daniel P. Berrangé wrote:
> Substitute in the @sysconfigdir@ value instead of /etc.
> 
> Signed-off-by: Daniel P. Berrangé <berrange at redhat.com>
> ---
>  src/locking/Makefile.inc.am         |  1 +
>  src/locking/virtlockd.service.in    |  2 +-
>  src/logging/Makefile.inc.am         |  1 +
>  src/logging/virtlogd.service.in     |  2 +-
>  src/remote/Makefile.inc.am          |  8 ++++++++
>  src/remote/libvirtd-admin.socket.in |  2 +-
>  src/remote/libvirtd-ro.socket.in    |  2 +-
>  src/remote/libvirtd-tcp.socket.in   |  2 +-
>  src/remote/libvirtd-tls.socket.in   |  2 +-
>  src/remote/libvirtd.aug.in          |  4 ++--
>  src/remote/libvirtd.conf.in         | 18 +++++++++---------
>  src/remote/libvirtd.service.in      |  2 +-
>  src/remote/libvirtd.socket.in       |  2 +-
>  src/remote/test_libvirtd.aug.in     |  8 ++++----
>  tools/libvirt-guests.service.in     |  2 +-
>  15 files changed, 34 insertions(+), 24 deletions(-)

After this patch, and indeed even after the entire series has been
applied, there are still a few instances left:

  src/locking/libvirt_lockd.aug:(* /etc/libvirt/qemu-lockd.conf *)
  src/locking/libvirt_lockd.aug:   let filter = incl "/etc/libvirt/qemu-lockd.conf"
  src/locking/libvirt_sanlock.aug:(* /etc/libvirt/qemu-sanlock.conf *)
  src/locking/libvirt_sanlock.aug:   let filter = incl "/etc/libvirt/qemu-sanlock.conf"
  src/locking/virtlockd.aug:(* /etc/libvirt/virtlockd.conf *)
  src/locking/virtlockd.aug:   let filter = incl "/etc/libvirt/virtlockd.conf"
  src/logging/virtlogd.aug:(* /etc/libvirt/virtlogd.conf *)
  src/logging/virtlogd.aug:   let filter = incl "/etc/libvirt/virtlogd.conf"
  src/remote/libvirtd.sasl:keytab: /etc/libvirt/krb5.tab
  src/remote/libvirtd.sasl:#sasldb_path: /etc/libvirt/passwd.db
  src/remote/libvirtd.sysconf:#KRB5_KTNAME=/etc/libvirt/krb5.tab
  src/remote/libvirtd.sysconf:# is enabled in /etc/libvirt/qemu.conf

Addressing the first bunch would also mean aligning virtlogd and
virtlockd with the rest of the daemons, which is a good thing.

Ideally we'd switch those files to being generated in a perparatory
patch and then update them along with everything else in this one,
but if you prefer doing it as a follow-up given how big this series
is already I'm okay with that as well.

In any case

  Reviewed-by: Andrea Bolognani <abologna at redhat.com>

for these changes.

-- 
Andrea Bolognani / Red Hat / Virtualization