QML profiler service sends memory events in wrong order

Consider:

iimport QtQuick 2.1
import QtQuick.Window 2.0

Window {
    width: 800
    height: 600
    visible: true

    Component.onCompleted: {
        var dictionaryTable = {};
        for (var j = 0; j < 646900; ++j) {

            var word = "a" + j.toString()
            dictionaryTable[word] = null;
        }
    }
}

Run this through the QML profiler and notice that all the memory usage events are aggregated into one. This is because all the memory usage events are sent in one block from QV4ProfilerAdapter::finalizeMessages() when the batch overflows. When appending memory events we need to check if there is a next call event and if so, only append memory events until that one.