Ability to deserialize TABLE_MAP and ROTATE events to byte arrays

[thomasdziedzic]

Related to #28 since the library depends on those being deserialized to a specific type.

[shyiko]

I think you can try something similar to:

public class CustomRotateEventData extends RotateEventData {

    private byte[] data;

    public byte[] getData() {
        return data;
    }

    public void setData(byte[] data) {
        this.data = data;
    }
}

public class CustomRotateEventDataDeserializer extends RotateEventDataDeserializer {

    @Override
    public RotateEventData deserialize(ByteArrayInputStream inputStream) throws IOException {
        byte[] data = inputStream.read(inputStream.available());
        RotateEventData eventData = super.deserialize(new ByteArrayInputStream(data));
        CustomRotateEventData result = new CustomRotateEventData();
        result.setData(data);
        result.setBinlogPosition(eventData.getBinlogPosition());
        result.setBinlogFilename(eventData.getBinlogFilename());
        return result;
    }
}

eventDeserializer.setEventDataDeserializer(EventType.ROTATE, new CustomRotateEventDataDeserializer());

Obviously, the same has to be done for TableMapEventData.

[thomasdziedzic]

This works, though I don't think it's ideal.

If I were to go with this solution, I would have to special case fetching byte data for all rotate and table_map events.

I think another solution could be to attach a getData method to the EventDataDeserializer class so that we can have a common interface for accessing byte data for all event types.

[shyiko]

Oh, I know it's far from being ideal. The workaround above I literally wrote while driving back home :) It's something that you can use now and not wait for me to change the internal flow (which is on its way).

[shyiko]

Fix available in 0.1.2-SNAPSHOT. Starting from 0417068 EventDataDeserializer can be specified regardless of the event type (= no special treatment for TABLE_MAP/ROTATE).

[shyiko]

0.1.2 should become available through Maven Central in the next 24h.

[thomasdziedzic]

@shyiko oh man, thanks for the quick implementation. I didn't want to sound ungrateful in my original post about your workaround. I just wanted to point out that it would be not ideal. Anyways, it seems like you went the extra step for this issue! cheers! 🍺

[shyiko]

I didn't want to sound...

Not at all, man. I truly appreciate your feedback. 🍻

[chenyun]

@shyiko Thanks a lot, after reading #29 #28 , I understand your design. 🍻