QML Property Changing Outside of Evaluation?

fcarney

                    property var def_g_segment: ({min:0, max:0, color:"white", inBkg:false})
                    property QtObject g_seg_temp: graph_model.getGraphSeg(index, graph_delegaterect.segment)
                    readonly property var g_segment: {
                        console.log("g_segment eval")
                        console.log("g_seg_temp", g_seg_temp)
                        if(g_seg_temp){
                            return g_seg_temp
                        }

                        return def_g_segment
                    }
                    onG_segmentChanged: {
                        if(1){
                            console.log("g_segment changed")
                            console.log(index, graph_model.numGraphers())
                            console.log(typeof(g_segment))
                            console.log(g_segment)
                        }
                    }

output:

qml: g_segment changed
qml: 0 6
qml: object
qml: null
qml: g_segment eval
qml: g_seg_temp null
qml: g_segment changed
qml: 0 6
qml: object
qml: [object Object]
qml: g_segment changed
qml: 0 6
qml: object
qml: null
qml: g_segment eval
qml: g_seg_temp null
qml: g_segment changed
qml: 0 6
qml: object
qml: [object Object]
qml: g_segment changed
qml: 0 6
qml: object
qml: null
qml: g_segment eval
qml: g_seg_temp null
qml: g_segment changed
qml: 0 6
qml: object
qml: [object Object]

graph_model.getGraphSeg can return null. So I am accounting for that in the evaluation of g_segment. However, I am seeing null come through under certain conditions. I want to detect null and use a dummy object instead. I am very confused as to how this is even possible. It seems like it is fighting itself somehow. I made it readonly on the property to ensure something else wasn't causing this. I also searched the code to find it, but just did not see anything. I need to try and mock this up in a test program.

Normally the text it prints is like this:

qml: g_segment eval
qml: g_seg_temp GraphSegment(0xd1b3b0)
qml: g_segment changed
qml: 0 6
qml: object
qml: GraphSegment(0xd1b3b0)

I have objects that are being updated affecting the index and segment values.

The graph_model is an old model that probably needs to be changed to a QAbstractListModel. That might be the real solution here. Just seems weird that the eval would be happening after a change signal. Unless the return value causes the signal to fire before the text is printed? I just don't know.