开发者

Immutable views of mutable types

开发者 https://www.devze.com 2023-01-25 01:14 出处:网络
I have a project where I need to construct a fair amount of configuration data before I can execute a process.During the configuration stage, it\'s very convenient to have the data as mutable.However,

I have a project where I need to construct a fair amount of configuration data before I can execute a process. During the configuration stage, it's very convenient to have the data as mutable. However, once configuration has been completed, I'd like to pass an immutable view of that data to the functional process, as that process will rely on configuration immutability for many of its computations (for instance, the ability to pre-compute things based on initial configuration.) I've come up with a possible solution using interfaces to expose a read-only view, but I'd like to know if anybody has encountered problems with this type of approach or if there are other recommendations for how to solve this problem.

One example of the pattern I'm currently using:

public interface IConfiguration
{
    string Version { get; }

    string VersionTag开发者_高级运维 { get; }

    IEnumerable<IDeviceDescriptor> Devices { get; }

    IEnumerable<ICommandDescriptor> Commands { get; }
}

[DataContract]
public sealed class Configuration : IConfiguration
{
    [DataMember]
    public string Version { get; set; }

    [DataMember]
    public string VersionTag { get; set; }

    [DataMember]
    public List<DeviceDescriptor> Devices { get; private set; }

    [DataMember]
    public List<CommandDescriptor> Commands { get; private set; }

    IEnumerable<IDeviceDescriptor> IConfiguration.Devices
    {
        get { return Devices.Cast<IDeviceDescriptor>(); }
    }

    IEnumerable<ICommandDescriptor> IConfiguration.Commands
    {
        get { return Commands.Cast<ICommandDescriptor>(); }
    }

    public Configuration()
    {
        Devices = new List<DeviceDescriptor>();
        Commands = new List<CommandDescriptor>();
    }
}

EDIT

Based on input from Mr. Lippert and cdhowie, I put together the following (removed some properties to simplify):

[DataContract]
public sealed class Configuration
{
    private const string InstanceFrozen = "Instance is frozen";

    private Data _data = new Data();
    private bool _frozen;

    [DataMember]
    public string Version
    {
        get { return _data.Version; }
        set
        {
            if (_frozen) throw new InvalidOperationException(InstanceFrozen);
            _data.Version = value;
        }
    }

    [DataMember]
    public IList<DeviceDescriptor> Devices
    {
        get { return _data.Devices; }
        private set { _data.Devices.AddRange(value); }
    }

    public IConfiguration Freeze()
    {
        if (!_frozen)
        {
            _frozen = true;
            _data.Devices.Freeze();
            foreach (var device in _data.Devices)
                device.Freeze();
        }
        return _data;
    }

    [OnDeserializing]
    private void OnDeserializing(StreamingContext context)
    {
        _data = new Data();
    }

    private sealed class Data : IConfiguration
    {
        private readonly FreezableList<DeviceDescriptor> _devices = new FreezableList<DeviceDescriptor>();

        public string Version { get; set; }

        public FreezableList<DeviceDescriptor> Devices
        {
            get { return _devices; }
        }

        IEnumerable<IDeviceDescriptor> IConfiguration.Devices
        {
            get { return _devices.Select(d => d.Freeze()); }
        }
    }
}

FreezableList<T> is, as you would expect, a freezable implementation of IList<T>. This gains insulation benefits, at the cost of some additional complexity.


The approach you describe works great if the "client" (the consumer of the interface) and the "server" (the provider of the class) have a mutual agreement that:

  • the client will be polite and not try to take advantage of the implementation details of the server
  • the server will be polite and not mutate the object after the client has a reference to it.

If you do not have a good working relationship between the people writing the client and the people writing the server then things go pear-shaped quickly. A rude client can of course "cast away" the immutability by casting to the public Configuration type. A rude server can hand out an immutable view and then mutate the object when the client least expects it.

A nice approach is to prevent the client from ever seeing the mutable type:

public interface IReadOnly { ... }
public abstract class Frobber : IReadOnly
{
    private Frobber() {}
    public class sealed FrobBuilder
    {
        private bool valid = true;
        private RealFrobber real = new RealFrobber();
        public void Mutate(...) { if (!valid) throw ... }
        public IReadOnly Complete { valid = false; return real; }
    }
    private sealed class RealFrobber : Frobber { ... }
}

Now if you want to create and mutate a Frobber, you can make a Frobber.FrobBuilder. When you're done your mutations, you call Complete and get a read-only interface. (And then the builder becomes invalid.) Since all the mutability implementation details are hidden in a private nested class, you can't "cast away" the IReadOnly interface to RealFrobber, only to Frobber, which has no public methods!

Nor can the hostile client create their own Frobber, because Frobber is abstract and has a private constructor. The only way to make a Frobber is via the builder.


This will work, but "malicious" methods may try to cast an IConfiguration to a Configuration and thereby bypass your interface-imposed restrictions. If you're not worried about that then your approach will work fine.

I usually do something like this:

public class Foo {
    private bool frozen = false;

    private string something;

    public string Something {
        get { return something; }
        set {
            if (frozen)
                throw new InvalidOperationException("Object is frozen.");

            // validate value

            something = value;
        }
    }

    public void Freeze() {
        frozen = true;
    }
}

Alternatively, you could deep-clone your mutable classes into immutable classes.


Why can't you provide a separate immutable view of the object?

public class ImmutableConfiguration {
    private Configuration _config;
    public ImmutableConfiguration(Configuration config) { _config = config; }
    public string Version { get { return _config.Version; } }
}

or if you don't like the extra typing, make the set members internal rather than public - accessible within the assembly but not by clients of it?


I'm regularly working with a large, COM-based framework (ESRI's ArcGIS Engine) that handles modifications very similarly in some situations: there are the "default" IFoo interfaces for read-only access, and IFooEdit interfaces (where applicable) for modifications.

That framework is fairly well-known, and I'm not aware of any widespread complaints about this particular design decision behind it.

Finally, I think it's definitely worth some additional thought in deciding which "perspective" gets to be the default one: the read-only perspective or the full-access one. I would personally make the read-only view the default.


How about:

struct Readonly<T>
{
    private T _value;
    private bool _hasValue;

    public T Value
    {
        get
        {
            if (!_hasValue)
                throw new InvalidOperationException();
            return _value;
        }
        set
        {
            if (_hasValue)
                throw new InvalidOperationException();
            _value = value;
        }
    }
}


[DataContract]
public sealed class Configuration
{
    private Readonly<string> _version;

    [DataMember]
    public string Version
    {
        get { return _version.Value; }
        set { _version.Value = value; }
    }
}

I called it Readonly but I'm not sure that's the best name for it though.

0

精彩评论

暂无评论...
验证码 换一张
取 消