There are a few cases where we don't want to clear a potential
nm-generated/volatile flag, but only mark the connection as
unsaved.

Otherwise, we wrongly end up clearing these flags and the connection
is wrongly not NM_DEVICE_SYS_IFACE_STATE_EXTERNAL.

Fixes: 35dc6421de78cffd1761cfb6edf9e94233f99966

When building with assertions, they nm_assert() for the
type. Otherwise, they are identical to a C cast.

Also, where possible, don't cast at all, but adjust
the type instead.

Also, there were a few missing casts.

ifnet shall use the new_connection argument, not NM_CONNECTION(self).
Also, let the caller of the virtual function provide the right new_connection,
not having the virtual function figure that out.

The current behavior of update_unsaved is confusing. Give the argument
an enum with a name that describes better what's happening. Also, it
makes the uses grep-able.

Next we will use siphash24() instead of the glib version g_direct_hash() or
g_str_hash(). Hence, the "nm-utils/nm-hash-utils.h" header becomes very
fundamental and will be needed basically everywhere.

Instead of requiring the users to include them, let it be included via
"nm-default.h" header.

During write, it can regularly happen that the connection gets modified.
For example, keyfile never writes blobs as-is, it always writes the
blob to an external file, and replaces the certificate property with
a path.
Other reasons could be just bugs, where the reader and writer are not doing
a proper round trip (these cases should be fixed).

Refactor commit_changes(), to return the re-read connection to
the settings-connection class, and handle replacing the settings
there.

Also, prepare for another change. Sometimes we first call replace_settings()
followed by commit_changes(). It would be better to instead call commit_changes()
first, and only on success proceed with replace_settings(). Hence, commit_changes()
gets a new argument new_connection, that can be used to write another
connection to disk.

Don't delegate so much to the virtual function delete().

Don't delegate so much to the virtual function commit_changes().
Calling the callback is not the task of the virtual function,
because every implementation must do that.

There are some minor changes in behavior for ifnet, where we now
first setup the monitors and reload the parsers, before invoking
the callback.