'has_key' on Dictionaries is removed from Python3 in favor of 'in'.

Signed-off-by: Thomas Haller <thaller@redhat.com>

libnm mostly used GPtrArrays in its APIs, except that arrays of
connections were usually GSLists. Fix this and make them GPtrArrays
too (and rename nm_client_list_connections() to
nm_client_get_connections() to match everything else).

Make NMRemoteSettings internal and have NMClient wrap all of its APIs,
just like it does with NMManager.

Make synchronous APIs take GCancellables, and make asynchronous APIs
use GAsyncReadyCallbacks and have names ending in "_async", with
"_finish" functions to retrieve the results.

Also, make nm_client_activate_connection_finish(),
nm_client_add_and_activate_finish(), and
nm_remote_settings_add_connection_finish() be (transfer full) rather
than (transfer none), because the refcounting semantics become
slightly confusing in some edge cases otherwise.

Merge nm_remote_settings_add_connection() and
nm_remote_settings_add_connection_unsaved(), and likewise
nm_remote_connection_commit_changes() and
nm_remote_connection_commit_changes_unsaved(), by adding a boolean
flag to each saying whether to save to disk.

Signed-off-by: Jiří Klimeš <jklimes@redhat.com>

https://bugzilla.gnome.org/show_bug.cgi?id=736802

Port libnm-core/libnm to GDBus.

The NetworkManager daemon continues to use dbus-glib; the
previously-added connection hash/variant conversion methods are now
moved to NetworkManagerUtils (along with a few other utilities that
are now only needed by the daemon code).

Port the dbus-glib-based examples to GDBus.

Also, don't use libnm in them at all; there's not much point in
examples that use the D-Bus API directly if they're just going to fall
back to libnm for the hard stuff... (And also, this avoids the problem
that GDBus uses GVariant, while the libnm-core APIs currently still
use GHashTables.)

Also fix up some comment grammar and copyright style, and add emacs
modelines where missing.

Also rename the existing GDBus-based examples to have names ending in
"-gdbus", not "-GDBus", since there's no reason to gratuitously
capitalize here.

libnm functions that return GPtrArrays of objects had a rule that if
the array was empty, they would return NULL rather than a 0-length
array. As it turns out, this is just a nuisance to clients, since in
most places the code for the non-empty case would end up doing the
right thing for the empty case as well (and where it doesn't, we can
check "array->len == 0" just as easily as "array == NULL"). So just
return the 0-length array instead.