Question

We have a framework built up whose API is this:

T Deserialize<T>(string to deserialize)
{
    //  XmlSerializer(typeof(T))
    //  deserialize and return
}

We’ve been passing it things like:

[XmlRoot("apple")]
public class Apple
{
    [XmlElement("Id")]
    public int AppleId { get; set; }
}

this works great when we are returned an apple, however, sometimes we are getting “notAnApple”. So I will either get “apple” or “notAnApple”. When I get “notAnApple”, I get an exception.

As we have a lot of code built around this method, I would prefer not to go around changing things like the method signature. I do, however, have full flexibility into passing whatever class I like into this deserialization method.

Answer 1

Generally you “know” exactly what you’re serializing and deserializing. e.g. you may serialize like this:

serializer.Serialize(stream, new Apple());
serializer.Serialize(stream, new notAnApple());

Then you’d serialize like:

Apple apple = (Apple)deserialzer.Deserialize(stream);
notAnApple peach = (notAnApple)desserializer.Deserialize(stream);

You don’t normally randomly try to deserialize data to figure out what it is that you’re deserializing. Technically you already know because you have a specific file format or a specific stream format.