0

在实现通用拖放模型时，我遇到了以下问题。在对mimeData进行编码时，我想将拖动索引的原始行和列写入数据流。项目数据的序列化取决于程序员选择的特定数据结构，因此应该封装在单独的方法中。DragDropListModel 的子类对其模型项使用不同的数据结构只需覆盖此封装方法。

问题是 QDataStream 上的读/写是有状态的。例如，请看一下 decode_data()。将 QDataStream 对象传递给负责从流中解码下一项的方法，（可能）调用复制构造函数。现在 decode_item_from_stream() 使用数据流的副本。在反序列化过程中完成的读取改变了复制的数据流对象的状态，但 decode_data() 中原始对象的状态保持不变。这应该会导致 while 循环的下一次迭代崩溃。

因此我的两个问题是：

  1. 我的描述是准确的还是我误解了什么？

  2. python 是否有一种方法可以规避这些问题（其他语言可能通过引用传递来解决这个问题），同时保持方法彼此分开？

    class DragDropListModel(QtCore.QAbstractListModel):
    def __init__(self, tabText):
        super(DragDropListModel, self).__init__()
        self.model_data = []

    def data(self, index, role):
        item_data = self.model_data[index.row()]

        # DEBUG
        if role == QtCore.Qt.DisplayRole:
            return item_data[generic_tokens.Display]
        if role == QtCore.Qt.DecorationRole:
            return item_data[generic_tokens.Decoration]
        if role == QtCore.Qt.ToolTipRole:
            item_data[generic_tokens.ToolTip]
        if role == generic_roles.ItemDataRole:
            return item_data

    def rowCount(self, index):
        return len(self.model_data)

    def mimeTypes(self):
        """Returns the list of allowed MIME types"""
        return ["application/x-generic-drag-drop"]

    # Returns an object that contains serialized items of data corresponding to the list of indexes specified. 
    # The format used to describe the encoded data is obtained from the mimeTypes() function
    def mimeData(self, indexes:list):
        mimeData = QtCore.QMimeData()
        encodedData = self.encode_data(indexes)
        mimeData.setData("application/x-generic-drag-drop", encodedData)
        return mimeData

    def encode_data(self, indexes:list):
        byteArray = QtCore.QByteArray()
        dataStream = QtCore.QDataStream(byteArray, QtCore.QIODevice.WriteOnly)
        # implement writing to datastream 
        for i, index in enumerate(indexes):
            # encode index
            dataStream.writeInt16(index.row())
            dataStream.writeInt16(index.column())
            # encode item
            dataStream = self.encode_item_to_stream(
                item=index.data(generic_roles.ItemDataRole)
                stream = dataStream
            )

        return byteArray

    def encode_item_to_stream(self, item, stream:QtCore.QDataStream) -> QtCore.QDataStream:
        """can be overridden by subclasses of the generic drag drop model to allow for custom item serialization"""
        pass

    def decode_data(self, data:QtCore.QMimeData)-> list: 
        # decoding the dropped data
        encodedData = data.data("application/x-generic-drag-drop")
        stream = QtCore.QDataStream(encodedData, QtCore.QIODevice.ReadOnly)

        source_indexes = []
        decoded_items = []
        while not stream.atEnd():
            # decode index
            row = stream.readInt16()
            col = stream.readInt16()
            source_indexes.append((row, col))
            # decode item
            decoded_items.append(stream)

        return indexes, decoded_items

    def decode_item_from_stream(self, stream:QtCore.QDataStream):
        """can be overridden by subclasses of the generic drag drop model to allow for custom item serialization"""
        pass
4

0 回答 0