Don't trigger udev if socket doesn't exist

On rpm-ostree systems, we don't want scriptlets to affect the running
system because a major part of the value is "background updates".
Scriptlets are run in a containerized environment where e.g. udev is not
available.

Add a check for the udev socket before triggering it to handle this.
This also helps the container use case.

This doesn't break rpm-ostree strictly, because it uses `|| :`, but it
still spams error messages during the compose. I kept the `|| :` to be
safe, but it's likely fine to remove them now.

See also: https://bugzilla.redhat.com/show_bug.cgi?id=1352154
See also: https://src.fedoraproject.org/rpms/udisks2/pull-request/3
See also: https://github.com/coreos/fedora-coreos-tracker/issues/703

https://src.fedoraproject.org/rpms/NetworkManager/pull-request/6

NetworkManager.spec

@@ -856,8 +856,12 @@ fi
 
 
 %post
+# skip triggering if udevd isn't even accessible, e.g. containers or
+# rpm-ostree-based systems
+if [ -S /run/udev/control ]; then
     /usr/bin/udevadm control --reload-rules || :
     /usr/bin/udevadm trigger --subsystem-match=net || :
+fi
 %if %{with firewalld_zone}
 %firewalld_reload
 %endif