[nemo-qml-plugin-dbus] add QDBusVariant type to typedCall marshaling

[CODeRUS]

No description provided.

[monich]

What's 'z'? Can't find it in any spec.

[CODeRUS]

Actually i don't know what single letter to assign for this type

[deztructor]

QDbusVariant is just a superset of QVariant? Do you need a specialization for it?

[CODeRUS]

I wouldn't make a PR if it will work with just QVariant. QDBusVariant is necessary to have.

[deztructor]

@CODeRUS Can you explain any use case?

[CODeRUS]

    DBusInterface {
        id: mceRequestIface
        service: 'com.nokia.mce'
        path: '/com/nokia/mce/request'
        iface: 'com.nokia.mce.request'
        bus: DBus.SystemBus

        function setValue(key, value) {
            typedCall('set_config', [{"type":"s", "value":key}, {"type":"z", "value":value}])
        }

        function getValue(key) {
            typedCall('get_config', [{"type":"s", "value":key}], function (value) {
                var temp = values
                temp[key] = value
                values = temp
            })
        }

        Component.onCompleted: {
            getValue(key_threshold_value)
            getValue(key_powersave_enable)
            getValue(key_powersave_force)
        }
    }

sending set_config with anything except QDBusVariant is not allowed

[monich]

Should the 'v' case be updated instead of inventing a new letter?

[spiiroin]

To me the whole function feels wrong. Instead of assuming random js variable can be returned
as qvariant and then pushed to dbus message in sensible way - why not pass appendable dbus args thingy + the js value and return true/false depending on whether the conversion could be made.

[CODeRUS]

@monich can be if it don't break anything using QVariant

[thp]

How does this relate to #30 ? Maybe we can solve the issue with either of these fixes?

[CODeRUS]

@thp it's not related. Just both PR is about using QDBusVariant from qml side. One is for receiving QDBusVariant value and mine is for sending QDBuaVariant

[spiiroin]

After 2nd look... The existing 'v' case does not guarantee that dbus variant gets used = is broken. The 'z' one by coderus does -> drop the non-standard 'z' and use the logic to handle the 'v' case?

[thp]

This needs updating if we want to merge it.

[thp]

@CODeRUS Do you think we can get this rebase against the master and merged?

[spiiroin]

This should not be needed anymore since the variant handling got eventually fixed in #30

@CODeRUS: Do you agree that this pull req can be closed?