Dela via


Binding types reference guide

This document describes the list of attributes that you can use to annotate your API contract files to drive the binding and the code generated

Xamarin.iOS and Xamarin.Mac API contracts are written in C# mostly as interface definitions that define the way that Objective-C code is surfaced to C#. The process involves a mix of interface declarations plus some basic type definitions that the API contract might require. For an introduction to binding types, see our companion guide Binding Objective-C Libraries.

Type definitions

Syntax:

[BaseType (typeof (BTYPE))
interface MyType : [Protocol1, Protocol2] {
     IntPtr Constructor (string foo);
}

Every interface in your contract definition that has the [BaseType] attribute declares the base type for the generated object. In the above declaration a MyType class C# type will be generated that binds to an Objective-C type called MyType.

If you specify any types after the typename (in the sample above Protocol1 and Protocol2) using the interface inheritance syntax the contents of those interfaces will be inlined as if they had been part of the contract for MyType. The way that Xamarin.iOS surfaces that a type adopts a protocol is by inlining all of the methods and properties that were declared in the protocol into the type itself.

The following shows how the Objective-C declaration for UITextField would be defined in a Xamarin.iOS contract:

@interface UITextField : UIControl <UITextInput> {

}

Would be written like this as a C# API contract:

[BaseType (typeof (UIControl))]
interface UITextField : UITextInput {
}

You can control many other aspects of the code generation by applying other attributes to the interface as well as configuring the [BaseType] attribute.

Generating events

One feature of the Xamarin.iOS and Xamarin.Mac API design is that we map Objective-C delegate classes as C# events and callbacks. Users can choose in a per-instance basis whether they want to adopt the Objective-C programming pattern, by assigning to properties like Delegate an instance of a class that implements the various methods that the Objective-C runtime would call, or by choosing the C#-style events and properties.

Let us see one example of how to use the Objective-C model:

bool MakeDecision ()
{
    return true;
}

void Setup ()
{
     var scrollView = new UIScrollView (myRect);
     scrollView.Delegate = new MyScrollViewDelegate ();
     ...
}

class MyScrollViewDelegate : UIScrollViewDelegate {
    public override void Scrolled (UIScrollView scrollView)
    {
        Console.WriteLine ("Scrolled");
    }

    public override bool ShouldScrollToTop (UIScrollView scrollView)
    {
        return MakeDecision ();
    }
}

In the above example, you can see that we have chosen to overwrite two methods, one a notification that a scrolling event has taken place, and the second that is a callback that should return a boolean value instructing the scrollView whether it should scroll to the top or not.

The C# model allows the user of your library to listen to notifications using the C# event syntax or the property syntax to hook up callbacks that are expected to return values.

This is how the C# code for the same feature looks like using lambdas:

void Setup ()
{
    var scrollview = new UIScrollView (myRect);
    // Event connection, use += and multiple events can be connected
    scrollView.Scrolled += (sender, eventArgs) { Console.WriteLine ("Scrolled"); }

    // Property connection, use = only a single callback can be used
    scrollView.ShouldScrollToTop = (sv) => MakeDecision ();
}

Since events do not return values (they have a void return type) you can connect multiple copies. The ShouldScrollToTop is not an event, it is instead a property with the type UIScrollViewCondition which has this signature:

public delegate bool UIScrollViewCondition (UIScrollView scrollView);

It returns a bool value, in this case the lambda syntax allows us to just return the value from the MakeDecision function.

The binding generator supports generating events and properties that link a class like UIScrollView with its UIScrollViewDelegate (well call these the Model class), this is done by annotating your [BaseType] definition with the Events and Delegates parameters (described below). In addition to annotating the [BaseType] with those parameters it is necessary to inform the generator of a few more components.

For events that take more than one parameter (in Objective-C the convention is that the first parameter in a delegate class is the instance of the sender object) you must provide the name that you would like for the generated EventArgs class to be. This is done with the [EventArgs] attribute on the method declaration in your Model class. For example:

[BaseType (typeof (UINavigationControllerDelegate))]
[Model][Protocol]
public interface UIImagePickerControllerDelegate {
    [Export ("imagePickerController:didFinishPickingImage:editingInfo:"), EventArgs ("UIImagePickerImagePicked")]
    void FinishedPickingImage (UIImagePickerController picker, UIImage image, NSDictionary editingInfo);
}

The above declaration will generate a UIImagePickerImagePickedEventArgs class that derives from EventArgs and packs both parameters, the UIImage and the NSDictionary. The generator produces this:

public partial class UIImagePickerImagePickedEventArgs : EventArgs {
    public UIImagePickerImagePickedEventArgs (UIImage image, NSDictionary editingInfo);
    public UIImage Image { get; set; }
    public NSDictionary EditingInfo { get; set; }
}

It then exposes the following in the UIImagePickerController class:

public event EventHandler<UIImagePickerImagePickedEventArgs> FinishedPickingImage { add; remove; }

Model methods that return a value are bound differently. Those require both a name for the generated C# delegate (the signature for the method) and also a default value to return in case the user does not provide an implementation. For example, the ShouldScrollToTop definition is this:

[BaseType (typeof (NSObject))]
[Model][Protocol]
public interface UIScrollViewDelegate {
    [Export ("scrollViewShouldScrollToTop:"), DelegateName ("UIScrollViewCondition"), DefaultValue ("true")]
    bool ShouldScrollToTop (UIScrollView scrollView);
}

The above will create a UIScrollViewCondition delegate with the signature that was shown above, and if the user does not provide an implementation, the return value will be true.

In addition to the [DefaultValue] attribute, you can also use the [DefaultValueFromArgument] attribute that directs the generator to return the value of the specified parameter in the call or the [NoDefaultValue] parameter that instructs the generator that there is no default value.

BaseTypeAttribute

Syntax:

public class BaseTypeAttribute : Attribute {
        public BaseTypeAttribute (Type t);

        // Properties
        public Type BaseType { get; set; }
        public string Name { get; set; }
        public Type [] Events { get; set; }
        public string [] Delegates { get; set; }
        public string KeepRefUntil { get; set; }
}

BaseType.Name

You use the Name property to control the name that this type will bind to in the Objective-C world. This is typically used to give the C# type a name that is compliant with the .NET Framework Design Guidelines, but which maps to a name in Objective-C that does not follow that convention.

Example, in the following case we map the Objective-C NSURLConnection type to NSUrlConnection, as the .NET Framework Design Guidelines use "Url" instead of "URL":

[BaseType (typeof (NSObject), Name="NSURLConnection")]
interface NSUrlConnection {
}

The specified name is used as the value for the generated [Register] attribute in the binding. If Name is not specified, the type's short name is used as the value for the [Register] attribute in the generated output.

BaseType.Events and BaseType.Delegates

These properties are used to drive the generation of C#-style events in the generated classes. They are used to link a given class with its Objective-C delegate class. You will encounter many cases where a class uses a delegate class to send notifications and events. For example a BarcodeScanner would have a companion BardodeScannerDelegate class. The BarcodeScanner class would typically have a Delegate property that you would assign an instance of BarcodeScannerDelegate to, while this works, you might want to expose to your users a C#-like style event interface, and in those cases you would use the Events and Delegates properties of the [BaseType] attribute.

These properties are always set together and must have the same number of elements and be kept in sync. The Delegates array contains one string for each weakly-typed delegate that you want to wrap, and the Events array contains one type for each type that you want to associate with it.

[BaseType (typeof (NSObject),
           Delegates=new string [] { "WeakDelegate" },
           Events=new Type [] {typeof(UIAccelerometerDelegate)})]
public interface UIAccelerometer {
}

[BaseType (typeof (NSObject))]
[Model][Protocol]
public interface UIAccelerometerDelegate {
}

BaseType.KeepRefUntil

If you apply this attribute when new instances of this class are created, the instance of that object will be kept around until the method referenced by the KeepRefUntil has been invoked. This is useful to improve the usability of your APIs, when you do not want your user to keep a reference to an object around to use your code. The value of this property is the name of a method in the Delegate class, so you must use this in combination with the Events and Delegates properties as well.

The following example show how this is used by UIActionSheet in Xamarin.iOS:

[BaseType (typeof (NSObject), KeepRefUntil="Dismissed")]
[BaseType (typeof (UIView),
           KeepRefUntil="Dismissed",
           Delegates=new string [] { "WeakDelegate" },
           Events=new Type [] {typeof(UIActionSheetDelegate)})]
public interface UIActionSheet {
}

[BaseType (typeof (NSObject))]
[Model][Protocol]
public interface UIActionSheetDelegate {
    [Export ("actionSheet:didDismissWithButtonIndex:"), EventArgs ("UIButton")]
    void Dismissed (UIActionSheet actionSheet, nint buttonIndex);
}

DesignatedDefaultCtorAttribute

When this attribute is applied to the interface definition it will generate a [DesignatedInitializer] attribute on the default (generated) constructor, which maps to the init selector.

DisableDefaultCtorAttribute

When this attribute is applied to the interface definition it will prevent the generator from producing the default constructor.

Use this attribute when you need the object to be initialized with one of the other constructors in the class.

PrivateDefaultCtorAttribute

When this attribute is applied to the interface definition it will flag the default constructor as private. This means that you can still instantiate object of this class internally from your extension file, but it just wont be accessible to users of your class.

CategoryAttribute

Use this attribute on a type definition to bind Objective-C categories and to expose those as C# extension methods to mirror the way Objective-C exposes the functionality.

Categories are an Objective-C mechanism used to extend the set of methods and properties available in a class. In practice, they are used to either extend the functionality of a base class (for example NSObject) when a specific framework is linked in (for example UIKit), making their methods available, but only if the new framework is linked in. In some other cases, they are used to organize features in a class by functionality. They are similar in spirit to C# extension methods.

This is what a category would look like in Objective-C:

@interface UIView (MyUIViewExtension)
-(void) makeBackgroundRed;
@end

The above example is found on a library that would extend instances of UIView with the method makeBackgroundRed.

To bind those, you can use the [Category] attribute on an interface definition. When using the [Category] attribute, the meaning of the [BaseType] attribute changes from being used to specify the base class to extend, to being the type to extend.

The following shows how the UIView extensions are bound and turned into C# extension methods:

[BaseType (typeof (UIView))]
[Category]
interface MyUIViewExtension {
    [Export ("makeBackgroundRed")]
    void MakeBackgroundRed ();
}

The above will create a MyUIViewExtension a class that contains the MakeBackgroundRed extension method. This means that you can now call MakeBackgroundRed on any UIView subclass, giving you the same functionality you would get on Objective-C.

In some cases you will find static members inside categories like in the following example:

@interface FooObject (MyFooObjectExtension)
+ (BOOL)boolMethod:(NSRange *)range;
@end

This will lead to an incorrect Category C# interface definition:

[Category]
[BaseType (typeof (FooObject))]
interface FooObject_Extensions {

    // Incorrect Interface definition
    [Static]
    [Export ("boolMethod:")]
    bool BoolMethod (NSRange range);
}

This is incorrect because to use the BoolMethod extension you need an instance of FooObject but you are binding an ObjC static extension, this is a side effect due to the fact of how C# extension methods are implemented.

The only way to use the above definitions is by the following ugly code:

(null as FooObject).BoolMethod (range);

The recommendation to avoid this is to inline the BoolMethod definition inside the FooObject interface definition itself, this will allow you to call this extension like it is intended FooObject.BoolMethod (range).

[BaseType (typeof (NSObject))]
interface FooObject {

    [Static]
    [Export ("boolMethod:")]
    bool BoolMethod (NSRange range);
}

We will issue a warning (BI1117) whenever we find a [Static] member inside a [Category] definition. If you really want to have [Static] members inside your [Category] definitions you can silence the warning by using [Category (allowStaticMembers: true)] or by decorating either your member or [Category] interface definition with [Internal].

StaticAttribute

When this attribute is applied to a class it will just generate a static class, one that does not derive from NSObject, so the [BaseType] attribute is ignored. Static classes are used to host C public variables that you want to expose.

For example:

[Static]
interface CBAdvertisement {
    [Field ("CBAdvertisementDataServiceUUIDsKey")]
    NSString DataServiceUUIDsKey { get; }

Will generate a C# class with the following API:

public partial class CBAdvertisement  {
    public static NSString DataServiceUUIDsKey { get; }
}

Protocol/Model definitions

Models are typically used by protocol implementation. They differ in that the runtime will only register with Objective-C the methods that actually have been overwritten. Otherwise, the method will not be registered.

This in general means that when you subclass a class that has been flagged with the ModelAttribute, you should not call the base method. Calling that method will throw the following exception: Foundation.You_Should_Not_Call_base_In_This_Method. You are supposed to implement the entire behavior on your subclass for any methods you override.

AbstractAttribute

By default, members that are part of a protocol are not mandatory. This allows users to create a subclass of the Model object by merely deriving from the class in C# and overriding only the methods they care about. Sometimes the Objective-C contract requires that the user provides an implementation for this method (those are flagged with the @required directive in Objective-C). In those cases, you should flag those methods with the [Abstract] attribute.

The [Abstract] attribute can be applied to either methods or properties and causes the generator to flag the generated member as abstract and the class to be an abstract class.

The following is taken from Xamarin.iOS:

[BaseType (typeof (NSObject))]
[Model][Protocol]
public interface UITableViewDataSource {
    [Export ("tableView:numberOfRowsInSection:")]
    [Abstract]
    nint RowsInSection (UITableView tableView, nint section);
}

DefaultValueAttribute

Specifies the default value to be returned by a model method if the user does not provide a method for this particular method in the Model object

Syntax:

public class DefaultValueAttribute : Attribute {
        public DefaultValueAttribute (object o);
        public object Default { get; set; }
}

For example, in the following imaginary delegate class for a Camera class, we provide a ShouldUploadToServer which would be exposed as a property on the Camera class. If the user of the Camera class does not explicitly set a the value to a lambda that can respond true or false, the default value return in this case would be false, the value that we specified in the DefaultValue attribute:

[BaseType (typeof (NSObject))]
[Model][Protocol]
interface CameraDelegate {
    [Export ("camera:shouldPromptForAction:"), DefaultValue (false)]
    bool ShouldUploadToServer (Camera camera, CameraAction action);
}

If the user sets a handler in the imaginary class, then this value would be ignored:

var camera = new Camera ();
camera.ShouldUploadToServer = (camera, action) => return SomeDecision ();

See also: [NoDefaultValue], [DefaultValueFromArgument].

DefaultValueFromArgumentAttribute

Syntax:

public class DefaultValueFromArgumentAttribute : Attribute {
    public DefaultValueFromArgumentAttribute (string argument);
    public string Argument { get; }
}

This attribute when provided on a method that returns a value on a model class will instruct the generator to return the value of the specified parameter if the user did not provide their own method or lambda.

Example:

[BaseType (typeof (NSObject))]
[Model][Protocol]
public interface NSAnimationDelegate {
    [Export ("animation:valueForProgress:"), DelegateName ("NSAnimationProgress"), DefaultValueFromArgumentAttribute ("progress")]
    float ComputeAnimationCurve (NSAnimation animation, nfloat progress);
}

In the above case if the user of the NSAnimation class chose to use any of the C# events/properties, and did not set NSAnimation.ComputeAnimationCurve to a method or lambda, the return value would be the value passed in the progress parameter.

See also: [NoDefaultValue], [DefaultValue]

IgnoredInDelegateAttribute

Sometimes it makes sense not to expose an event or delegate property from a Model class into the host class so adding this attribute will instruct the generator to avoid the generation of any method decorated with it.

[BaseType (typeof (UINavigationControllerDelegate))]
[Model][Protocol]
public interface UIImagePickerControllerDelegate {
    [Export ("imagePickerController:didFinishPickingImage:editingInfo:"), EventArgs ("UIImagePickerImagePicked")]
    void FinishedPickingImage (UIImagePickerController picker, UIImage image, NSDictionary editingInfo);

    [Export ("imagePickerController:didFinishPickingImage:"), IgnoredInDelegate)] // No event generated for this method
    void FinishedPickingImage (UIImagePickerController picker, UIImage image);
}

DelegateNameAttribute

This attribute is used in Model methods that return values to set the name of the delegate signature to use.

Example:

[BaseType (typeof (NSObject))]
[Model][Protocol]
public interface NSAnimationDelegate {
    [Export ("animation:valueForProgress:"), DelegateName ("NSAnimationProgress"), DefaultValueFromArgumentAttribute ("progress")]
    float ComputeAnimationCurve (NSAnimation animation, float progress);
}

With the above definition, the generator will produce the following public declaration:

public delegate float NSAnimationProgress (MonoMac.AppKit.NSAnimation animation, float progress);

DelegateApiNameAttribute

This attribute is used to allow the generator to change the name of the property generated in the host class. Sometimes it is useful when the name of the FooDelegate class method makes sense for the Delegate class, but would look odd in the host class as a property.

Also this is really useful (and needed) when you have two or more overload methods that makes sense to keep them named as is in the FooDelegate class but you want to expose them in the host class with a better given name.

Example:

[BaseType (typeof (NSObject))]
[Model][Protocol]
public interface NSAnimationDelegate {
    [Export ("animation:valueForProgress:"), DelegateApiName ("ComputeAnimationCurve"), DelegateName ("Func<NSAnimation, float, float>"), DefaultValueFromArgument ("progress")]
    float GetValueForProgress (NSAnimation animation, float progress);
}

With the above definition, the generator will produce the following public declaration in the host class:

public Func<NSAnimation, float, float> ComputeAnimationCurve { get; set; }

EventArgsAttribute

For events that take more than one parameter (in Objective-C the convention is that the first parameter in a delegate class is the instance of the sender object) you must provide the name that you would like for the generated EventArgs class to be. This is done with the [EventArgs] attribute on the method declaration in your Model class.

For example:

[BaseType (typeof (UINavigationControllerDelegate))]
[Model][Protocol]
public interface UIImagePickerControllerDelegate {
    [Export ("imagePickerController:didFinishPickingImage:editingInfo:"), EventArgs ("UIImagePickerImagePicked")]
    void FinishedPickingImage (UIImagePickerController picker, UIImage image, NSDictionary editingInfo);
}

The above declaration will generate a UIImagePickerImagePickedEventArgs class that derives from EventArgs and packs both parameters, the UIImage and the NSDictionary. The generator produces this:

public partial class UIImagePickerImagePickedEventArgs : EventArgs {
    public UIImagePickerImagePickedEventArgs (UIImage image, NSDictionary editingInfo);
    public UIImage Image { get; set; }
    public NSDictionary EditingInfo { get; set; }
}

It then exposes the following in the UIImagePickerController class:

public event EventHandler<UIImagePickerImagePickedEventArgs> FinishedPickingImage { add; remove; }

EventNameAttribute

This attribute is used to allow the generator to change the name of an event or property generated in the class. Sometimes it is useful when the name of the Model class method makes sense for the model class, but would look odd in the originating class as an event or property.

For example, the UIWebView uses the following bit from the UIWebViewDelegate:

[Export ("webViewDidFinishLoad:"), EventArgs ("UIWebView"), EventName ("LoadFinished")]
void LoadingFinished (UIWebView webView);

The above exposes LoadingFinished as the method in the UIWebViewDelegate, but LoadFinished as the event to hook up to in a UIWebView:

var webView = new UIWebView (...);
webView.LoadFinished += delegate { Console.WriteLine ("done!"); }

ModelAttribute

When you apply the [Model] attribute to a type definition in your contract API, the runtime will generate special code that will only surface invocations to methods in the class if the user has overwritten a method in the class. This attribute is typically applied to all APIs that wrap an Objective-C delegate class.

The runtime will also generate an Objective-C class that matches the name of the corresponding protocol.

It's possible to customize the name of the Objective-C class in two ways:

  1. Setting AutoGeneratedName = true:

    [Model (AutoGeneratedName = true)]
    

    This will make the runtime generate an unique name for the Objective-C type. The name is currently based on the assembly name and the full name of the model's type (this may change in the future).

  2. Explicitly specifying the name:

    [Model (Name = "CustomName")]
    

It's recommended to use AutoGeneratedName = true. In .NET, the name is always generated (unless it's explicitly specified as in 2. above), and the AutoGeneratedName property does not exist anymore.

NoDefaultValueAttribute

Specifies that the method on the model does not provide a default return value.

This works with the Objective-C runtime by responding false to the Objective-C runtime request to determine if the specified selector is implemented in this class.

[BaseType (typeof (NSObject))]
[Model][Protocol]
interface CameraDelegate {
    [Export ("shouldDisplayPopup"), NoDefaultValue]
    bool ShouldUploadToServer ();
}

See also: [DefaultValue], [DefaultValueFromArgument]

Protocols

The Objective-C protocol concept does not really exist in C#. Protocols are similar to C# interfaces but they differ in that not all of the methods and properties declared in a protocol must be implemented by the class that adopts it. Instead some of the methods and properties are optional.

Some protocols are generally used as Model classes, those should be bound using the [Model] attribute.

[BaseType (typeof (NSObject))]
[Model, Protocol]
interface MyProtocol {
    // Use [Abstract] when the method is defined in the @required section
    // of the protocol definition in Objective-C
    [Abstract]
    [Export ("say:")]
    void Say (string msg);

    [Export ("listen")]
    void Listen ();
}

Starting with Xamarin.iOS 7.0 a new and improved protocol binding functionality has been incorporated. Any definition that contains the [Protocol] attribute will actually generate three supporting classes that vastly improve the way that you consume protocols:

// Full method implementation, contains all methods
class MyProtocol : IMyProtocol {
    public void Say (string msg);
    public void Listen (string msg);
}

// Interface that contains only the required methods
interface IMyProtocol: INativeObject, IDisposable {
    [Export ("say:")]
    void Say (string msg);
}

// Extension methods
static class IMyProtocol_Extensions {
    public static void Optional (this IMyProtocol this, string msg);
    }
}

The class implementation provides a complete abstract class that you can override individual methods of and get full type safety. But due to C# not supporting multiple inheritance, there are scenarios where you might require a different base class, but still want to implement an interface.

This is where the generated interface definition comes in. It is an interface that has all the required methods from the protocol. This allows developers that want to implement your protocol to merely implement the interface. The runtime will automatically register the type as adopting the protocol.

Notice that the interface only lists the required methods and does expose the optional methods. This means that classes that adopt the protocol will get full type checking for the required methods, but will have to resort to weak typing (manually using Export attributes and matching the signature) for the optional protocol methods.

To make it convenient to consume an API that uses protocols, the binding tool also will produce an extensions method class that exposes all of the optional methods. This means that as long as you are consuming an API, you will be able to treat protocols as having all the methods.

If you want to use the protocol definitions in your API, you will need to write skeleton empty interfaces in your API definition. If you want to use the MyProtocol in an API, you would need to do this:

[BaseType (typeof (NSObject))]
[Model, Protocol]
interface MyProtocol {
    // Use [Abstract] when the method is defined in the @required section
    // of the protocol definition in Objective-C
    [Abstract]
    [Export ("say:")]
    void Say (string msg);

    [Export ("listen")]
    void Listen ();
}

interface IMyProtocol {}

[BaseType (typeof(NSObject))]
interface MyTool {
    [Export ("getProtocol")]
    IMyProtocol GetProtocol ();
}

The above is needed because at binding time the IMyProtocol would not exist, that is why you need to provide an empty interface.

Adopting protocol-generated interfaces

Whenever you implement one of the interfaces generated for the protocols, like this:

class MyDelegate : NSObject, IUITableViewDelegate {
    nint IUITableViewDelegate.GetRowHeight (nint row) {
        return 1;
    }
}

The implementation for the required interface methods gets exported with the proper name, so it is equivalent to this:

class MyDelegate : NSObject, IUITableViewDelegate {
    [Export ("getRowHeight:")]
    nint IUITableViewDelegate.GetRowHeight (nint row) {
        return 1;
    }
}

This will work for all required protocol members, but there is a special case with optional selectors to be aware of.

Optional protocol members are treated identically when using the base class:

public class UrlSessionDelegate : NSUrlSessionDownloadDelegate {
	public override void DidWriteData (NSUrlSession session, NSUrlSessionDownloadTask downloadTask, long bytesWritten, long totalBytesWritten, long totalBytesExpectedToWrite)

but when using the protocol interface it is required to add the [Export]. The IDE will add it via autocomplete when you add it starting with override.

public class UrlSessionDelegate : NSObject, INSUrlSessionDownloadDelegate {
	[Export ("URLSession:downloadTask:didWriteData:totalBytesWritten:totalBytesExpectedToWrite:")]
	public void DidWriteData (NSUrlSession session, NSUrlSessionDownloadTask downloadTask, long bytesWritten, long totalBytesWritten, long totalBytesExpectedToWrite)

There is a slight behavior difference between the two at runtime:

  • Users of the base class (NSUrlSessionDownloadDelegate in example) provides all required and optional selectors, returning reasonable default values.
  • Users of the interface (INSUrlSessionDownloadDelegate in example) only respond to the exact selectors provided.

Some rare classes can behave differently here. In almost all cases however it is safe to use either.

Protocol inlining

While you bind existing Objective-C types that have been declared as adopting a protocol, you will want to inline the protocol directly. To do this, merely declare your protocol as an interface without any [BaseType] attribute and list the protocol in the list of base interfaces for your interface.

Example:

interface SpeakProtocol {
    [Export ("say:")]
    void Say (string msg);
}

[BaseType (typeof (NSObject))]
interface Robot : SpeakProtocol {
    [Export ("awake")]
    bool Awake { get; set; }
}

Member definitions

The attributes in this section are applied to individual members of a type: properties and method declarations.

AlignAttribute

Used to specify the alignment value for property return types. Certain properties take pointers to addresses that must be aligned at certain boundaries (in Xamarin.iOS this happens for example with some GLKBaseEffect properties that must be 16-byte aligned). You can use this property to decorate the getter, and use the alignment value. This is typically used with the OpenTK.Vector4 and OpenTK.Matrix4 types when integrated with Objective-C APIs.

Example:

public interface GLKBaseEffect {
    [Export ("constantColor")]
    Vector4 ConstantColor { [Align (16)] get; set;  }
}

AppearanceAttribute

The [Appearance] attribute is limited to iOS 5, where the Appearance manager was introduced.

The [Appearance] attribute can be applied to any method or property that participate in the UIAppearance framework. When this attribute is applied to a method or property in a class, it will direct the binding generator to create a strongly-typed appearance class that is used to style all the instances of this class, or the instances that match certain criteria.

Example:

public interface UIToolbar {
    [Export ("setBackgroundImage:forToolbarPosition:barMetrics:")]
    [Appearance]
    void SetBackgroundImage (UIImage backgroundImage, UIToolbarPosition position, UIBarMetrics barMetrics);

    [Export ("backgroundImageForToolbarPosition:barMetrics:")]
    [Appearance]
    UIImage GetBackgroundImage (UIToolbarPosition position, UIBarMetrics barMetrics);
}

The above would generate the following code in UIToolbar:

public partial class UIToolbar {
    public partial class UIToolbarAppearance : UIView.UIViewAppearance {
        public virtual void SetBackgroundImage (UIImage backgroundImage, UIToolbarPosition position, UIBarMetrics barMetrics);
        public virtual UIImage GetBackgroundImage (UIToolbarPosition position, UIBarMetrics barMetrics)
    }
    public static new UIToolbarAppearance Appearance { get; }
    public static new UIToolbarAppearance AppearanceWhenContainedIn (params Type [] containers);
}

AutoReleaseAttribute (Xamarin.iOS 5.4)

Use the [AutoReleaseAttribute] on methods and properties to wrap the method invocation to the method in an NSAutoReleasePool.

In Objective-C there are some methods that return values that are added to the default NSAutoReleasePool. By default, these would go to your thread NSAutoReleasePool, but since Xamarin.iOS also keeps a reference to your objects as long as the managed object lives, you might not want to keep an extra reference in the NSAutoReleasePool which will only get drained until your thread returns control to the next thread, or you go back to the main loop.

This attribute is applied for example on heavy properties (for example UIImage.FromFile) that returns objects that have been added to the default NSAutoReleasePool. Without this attribute, the images would be retained as long as your thread did not return control to the main loop. Uf your thread was some sort of background downloader that is always alive and waiting for work, the images would never be released.

ForcedTypeAttribute

The [ForcedTypeAttribute] is used to enforce the creation of a managed type even if the returned unmanaged object does not match the type described in the binding definition.

This is useful when the type described in a header does not match the returned type of the native method, for example take the following Objective-C definition from NSURLSession:

- (NSURLSessionDownloadTask *)downloadTaskWithRequest:(NSURLRequest *)request

It clearly states that it will return an NSURLSessionDownloadTask instance, but yet it returns a NSURLSessionTask, which is a superclass and thus not convertible to NSURLSessionDownloadTask. Since we are in a type-safe context an InvalidCastException will happen.

To comply with the header description and avoid the InvalidCastException, the [ForcedTypeAttribute] is used.

[BaseType (typeof (NSObject), Name="NSURLSession")]
interface NSUrlSession {

    [Export ("downloadTaskWithRequest:")]
    [return: ForcedType]
    NSUrlSessionDownloadTask CreateDownloadTask (NSUrlRequest request);
}

The [ForcedTypeAttribute] also accepts a boolean value named Owns that is false by default [ForcedType (owns: true)]. The owns parameter is used to follow the Ownership Policy for Core Foundation objects.

The [ForcedTypeAttribute] is only valid on parameters, properties, and return value.

BindAsAttribute

The [BindAsAttribute] allows binding NSNumber, NSValue and NSString(enums) into more accurate C# types. The attribute can be used to create better, more accurate, .NET API over the native API.

You can decorate methods (on return value), parameters and properties with BindAs. The only restriction is that your member must not be inside a [Protocol] or [Model] interface.

For example:

[return: BindAs (typeof (bool?))]
[Export ("shouldDrawAt:")]
NSNumber ShouldDraw ([BindAs (typeof (CGRect))] NSValue rect);

Would output:

[Export ("shouldDrawAt:")]
bool? ShouldDraw (CGRect rect) { ... }

Internally we will do the bool? <-> NSNumber and CGRect <-> NSValue conversions.

The current supported encapsulation types are:

  • NSValue
  • NSNumber
  • NSString

NSValue

The following C# data types are supported to be encapsulated from/into NSValue:

  • CGAffineTransform
  • NSRange
  • CGVector
  • SCNMatrix4
  • CLLocationCoordinate2D
  • SCNVector3
  • SCNVector4
  • CGPoint / PointF
  • CGRect / RectangleF
  • CGSize / SizeF
  • UIEdgeInsets
  • UIOffset
  • MKCoordinateSpan
  • CMTimeRange
  • CMTime
  • CMTimeMapping
  • CATransform3D

NSNumber

The following C# data types are supported to be encapsulated from/into NSNumber:

  • bool
  • byte
  • double
  • float
  • short
  • int
  • long
  • sbyte
  • ushort
  • uint
  • ulong
  • nfloat
  • nint
  • nuint
  • Enums

NSString

[BindAs] works in conjuntion with enums backed by a NSString constant so you can create better .NET API, for example:

[BindAs (typeof (CAScroll))]
[Export ("supportedScrollMode")]
NSString SupportedScrollMode { get; set; }

Would output:

[Export ("supportedScrollMode")]
CAScroll SupportedScrollMode { get; set; }

We will handle the enum <-> NSString conversion only if the provided enum type to [BindAs] is backed by a NSString constant.

Arrays

[BindAs] also supports arrays of any of the supported types, you can have the following API definition as an example:

[return: BindAs (typeof (CAScroll []))]
[Export ("getScrollModesAt:")]
NSString [] GetScrollModes ([BindAs (typeof (CGRect []))] NSValue [] rects);

Would output:

[Export ("getScrollModesAt:")]
CAScroll? [] GetScrollModes (CGRect [] rects) { ... }

The rects parameter will be encapsulated into a NSArray that contains an NSValue for each CGRect and in return you will get an array of CAScroll? which has been created using the values of the returned NSArray containing NSStrings.

BindAttribute

The [Bind] attribute has two uses one when applied to a method or property declaration, and another one when applied to the individual getter or setter in a property.

When used for a method or property, the effect of the [Bind] attribute is to generate a method that invokes the specified selector. But the resulting generated method is not decorated with the [Export] attribute, which means that it can not participate in method overriding. This is typically used in combination with the [Target] attribute for implementing Objective-C extension methods.

For example:

public interface UIView {
    [Bind ("drawAtPoint:withFont:")]
    SizeF DrawString ([Target] string str, CGPoint point, UIFont font);
}

When used in a getter or setter, the [Bind] attribute is used to alter the defaults inferred by the code generator when generating the getter and setter Objective-C selector names for a property. By default when you flag a property with the name fooBar, the generator would generate a fooBar export for the getter and setFooBar: for the setter. In a few cases, Objective-C does not follow this convention, usually they change the getter name to be isFooBar. You would use this attribute to inform the generator of this.

For example:

// Default behavior
[Export ("active")]
bool Active { get; set; }

// Custom naming with the Bind attribute
[Export ("visible")]
bool Visible { [Bind ("isVisible")] get; set; }

AsyncAttribute

Only available on Xamarin.iOS 6.3 and newer.

This attribute can be applied to methods that take a completion handler as their last argument.

You can use the [Async] attribute on methods whose last argument is a callback. When you apply this to a method, the binding generator will generate a version of that method with the suffix Async. If the callback takes no parameters, the return value will be a Task, if the callback takes a parameter, the result will be a Task<T>.

[Export ("upload:complete:")]
[Async]
void LoadFile (string file, NSAction complete)

The following will generate this async method:

Task LoadFileAsync (string file);

If the callback takes multiple parameters, you should set the ResultType or ResultTypeName to specify the desired name of the generated type which will hold all the properties.

delegate void OnComplete (string [] files, nint byteCount);

[Export ("upload:complete:")]
[Async (ResultTypeName="FileLoading")]
void LoadFiles (string file, OnComplete complete)

The following will generate this async method, where FileLoading contains properties to access both files and byteCount:

Task<FileLoading> LoadFile (string file);

If the last parameter of the callback is an NSError, then the generated Async method will check if the value is not null, and if that is the case, the generated async method will set the task exception.

[Export ("upload:onComplete:")]
[Async]
void Upload (string file, Action<string,NSError> onComplete);

The above generates the following async method:

Task<string> UploadAsync (string file);

And on error, the resulting Task will have the exception set to an NSErrorException that wraps the resulting NSError.

AsyncAttribute.ResultType

Use this property to specify the value for the returning Task object. This parameter takes an existing type, thus it needs to be defined in one of your core api definitions.

AsyncAttribute.ResultTypeName

Use this property to specify the value for the returning Task object. This parameter takes the name of your desired type name, the generator will produce a series of properties, one for each parameter that the callback takes.

AsyncAttribute.MethodName

Use this property to customize the name of the generated async methods. The default is to use the name of the method and append the text "Async", you can use this to change this default.

DesignatedInitializerAttribute

When this attribute is applied to a constructor it will generate the same [DesignatedInitializer] in the final platform assembly. This is to help the IDE indicate which constructor should be used in subclasses.

This should map to Objective-C/clang use of __attribute__((objc_designated_initializer)).

DisableZeroCopyAttribute

This attribute is applied to string parameters or string properties and instructs the code generator to not use the zero-copy string marshaling for this parameter, and instead create a new NSString instance from the C# string. This attribute is only required on strings if you instruct the generator to use zero-copy string marshaling using either the --zero-copy command line option or setting the assembly-level attribute ZeroCopyStringsAttribute.

This is necessary in cases where the property is declared in Objective-C to be a retain or assign property instead of a copy property. These typically happen in third-party libraries that have been wrongly "optimized" by developers. In general, retain or assign NSString properties are incorrect since NSMutableString or user-derived classes of NSString might alter the contents of the strings without the knowledge of the library code, subtly breaking the application. Typically this happens due to premature optimization.

The following shows two such properties in Objective-C:

@property(nonatomic,retain) NSString *name;
@property(nonatomic,assign) NSString *name2;

DisposeAttribute

When you apply the [DisposeAttribute] to a class, you provide a code snippet that will be added to the Dispose() method implementation of the class.

Since the Dispose method is automatically generated by the bgen tool, you need to use the [Dispose] attribute to inject some code in the generated Dispose method implementation.

For example:

[BaseType (typeof (NSObject))]
[Dispose ("if (OpenConnections > 0) CloseAllConnections ();")]
interface DatabaseConnection {
}

ExportAttribute

The [Export] attribute is used to flag a method or property to be exposed to the Objective-C runtime. This attribute is shared between the binding tool and the actual Xamarin.iOS and Xamarin.Mac runtimes. For methods, the parameter is passed verbatim to the generated code, for properties, a getter and setter Exports are generated based on the base declaration (see the section on the [BindAttribute] for information on how to alter the behavior of the binding tool).

Syntax:

public enum ArgumentSemantic {
    None, Assign, Copy, Retain.
}

[AttributeUsage (AttributeTargets.Method | AttributeTargets.Constructor | AttributeTargets.Property)]
public class ExportAttribute : Attribute {
    public ExportAttribute();
    public ExportAttribute (string selector);
    public ExportAttribute (string selector, ArgumentSemantic semantic);
    public string Selector { get; set; }
    public ArgumentSemantic ArgumentSemantic { get; set; }
}

The selector represents the name of the underlying Objective-C method or property that is being bound.

ExportAttribute.ArgumentSemantic

FieldAttribute

This attribute is used to expose a C global variable as a field that is loaded on demand and exposed to C# code. Usually this is required to get the values of constants that are defined in C or Objective-C and that could be either tokens used in some APIs, or whose values are opaque and must be used as-is by user code.

Syntax:

public class FieldAttribute : Attribute {
    public FieldAttribute (string symbolName);
    public FieldAttribute (string symbolName, string libraryName);
    public string SymbolName { get; set; }
    public string LibraryName { get; set; }
}

The symbolName is the C symbol to link with. By default this will be loaded from a library whose name is inferred from the namespace where the type is defined. If this is not the library where the symbol is looked up, you should pass the libraryName parameter. If you're linking a static library, use __Internal as the libraryName parameter.

The generated properties are always static.

Properties flagged with the Field attribute can be of the following types:

  • NSString
  • NSArray
  • nint / int / long
  • nuint / uint / ulong
  • nfloat / float
  • double
  • CGSize
  • System.IntPtr
  • Enums

Setters are not supported for enums backed by NSString constants, but they can be manually bound if needed.

Example:

[Static]
interface CameraEffects {
     [Field ("kCameraEffectsZoomFactorKey", "CameraLibrary")]
     NSString ZoomFactorKey { get; }
}

InternalAttribute

The [Internal] attribute can be applied to methods or properties and it has the effect of flagging the generated code with the internal C# keyword making the code only accessible to code in the generated assembly. This is typically used to hide APIs that are too low-level or provide a suboptimal public API that you want to improve upon or for APIs that are not supported by the generator and require some hand-coding.

When you design the binding, you would typically hide the method or property using this attribute and provide a different name for the method or property, and then on your C# complementary support file, you would add a strongly-typed wrapper that exposes the underlying functionality.

For example:

[Internal]
[Export ("setValue:forKey:")]
void _SetValueForKey (NSObject value, NSObject key);

[Internal]
[Export ("getValueForKey:")]
NSObject _GetValueForKey (NSObject key);

Then, in your supporting file, you could have some code like this:

public NSObject this [NSObject idx] {
    get {
        return _GetValueForKey (idx);
    }
    set {
        _SetValueForKey (value, idx);
    }
}

IsThreadStaticAttribute

This attribute flags the backing field for a property to be annotated with the .NET [ThreadStatic] attribute. This is useful if the field is a thread static variable.

MarshalNativeExceptions (Xamarin.iOS 6.0.6)

This attribute will make a method support native (Objective-C) exceptions. Instead of calling objc_msgSend directly, the invocation will go through a custom trampoline which catches ObjectiveC exceptions and marshals them into managed exceptions.

Currently only a few objc_msgSend signatures are supported (you will find out if a signature isn't supported when native linking of an app that uses the binding fails with a missing xamarin__objc_msgSend symbol), but more can be added at request.

NewAttribute

This attribute is applied to methods and properties to have the generator generate the new keyword in front of the declaration.

It is used to avoid compiler warnings when the same method or property name is introduced in a subclass that already existed in a base class.

NotificationAttribute

You can apply this attribute to fields to have the generator produce a strongly-typed helper Notifications class.

This attribute can be used without arguments for notifications that carry no payload, or you can specify a System.Type that references another interface in the API definition, typically with the name ending with "EventArgs". The generator will turn the interface into a class that subclasses EventArgs and will include all of the properties listed there. The [Export] attribute should be used in the EventArgs class to list the name of the key used to look up the Objective-C dictionary to fetch the value.

For example:

interface MyClass {
    [Notification]
    [Field ("MyClassDidStartNotification")]
    NSString DidStartNotification { get; }
}

The above code will generate a nested class MyClass.Notifications with the following methods:

public class MyClass {
   [..]
   public Notifications {
      public static NSObject ObserveDidStart (EventHandler<NSNotificationEventArgs> handler)
      public static NSObject ObserveDidStart (NSObject objectToObserve, EventHandler<NSNotificationEventArgs> handler)
   }
}

Users of your code can then easily subscribe to notifications posted to the NSDefaultCenter by using code like this:

var token = MyClass.Notifications.ObserverDidStart ((notification) => {
    Console.WriteLine ("Observed the 'DidStart' event!");
});

Or to set a specific object to observe. If you pass null to objectToObserve this method will behave just like its other peer.

var token = MyClass.Notifications.ObserverDidStart (objectToObserve, (notification) => {
    Console.WriteLine ("Observed the 'DidStart' event on objectToObserve!");
});

The returned value from ObserveDidStart can be used to easily stop receiving notifications, like this:

token.Dispose ();

Or you can call NSNotification.DefaultCenter.RemoveObserver and pass the token. If your notification contains parameters, you should specify a helper EventArgs interface, like this:

interface MyClass {
    [Notification (typeof (MyScreenChangedEventArgs)]
    [Field ("MyClassScreenChangedNotification")]
    NSString ScreenChangedNotification { get; }
}

// The helper EventArgs declaration
interface MyScreenChangedEventArgs {
    [Export ("ScreenXKey")]
    nint ScreenX { get; set; }

    [Export ("ScreenYKey")]
    nint ScreenY { get; set; }

    [Export ("DidGoOffKey")]
    [ProbePresence]
    bool DidGoOff { get; }
}

The above will generate a MyScreenChangedEventArgs class with the ScreenX and ScreenY properties that will fetch the data from the NSNotification.UserInfo dictionary using the keys ScreenXKey and ScreenYKey respectively and apply the proper conversions. The [ProbePresence] attribute is used for the generator to probe if the key is set in the UserInfo, instead of trying to extract the value. This is used for cases where the presence of the key is the value (typically for boolean values).

This allows you to write code like this:

var token = MyClass.NotificationsObserveScreenChanged ((notification) => {
    Console.WriteLine ("The new screen dimensions are {0},{1}", notification.ScreenX, notification.ScreenY);
});

In some cases, there is no constant associated with the value passed on the dictionary. Apple sometimes uses public symbol constants and sometimes uses string constants. By default the [Export] attribute in your provided EventArgs class will use the specified name as a public symbol to be looked up at runtime. If this is not the case, and instead it is supposed to be looked up as a string constant then pass the ArgumentSemantic.Assign value to the Export attribute.

New in Xamarin.iOS 8.4

Sometimes, notifications will begin life without any arguments, so the use of [Notification] without arguments is acceptable. But sometimes, parameters to the notification will be introduced. To support this scenario, the attribute can be applied more than once.

If you are developing a binding, and you want to avoid breaking existing user code, you would turn an existing notification from:

interface MyClass {
    [Notification]
    [Field ("MyClassScreenChangedNotification")]
    NSString ScreenChangedNotification { get; }
}

Into a version that lists the notification attribute twice, like this:

interface MyClass {
    [Notification]
    [Notification (typeof (MyScreenChangedEventArgs)]
    [Field ("MyClassScreenChangedNotification")]
    NSString ScreenChangedNotification { get; }
}

NullAllowedAttribute

When this is applied to a property it flags the property as allowing the value null to be assigned to it. This is only valid for reference types.

When this is applied to a parameter in a method signature it indicates that the specified parameter can be null and that no check should be performed for passing null values.

If the reference type does not have this attribute, the binding tool will generate a check for the value being assigned before passing it to Objective-C and will generate a check that will throw an ArgumentNullException if the value assigned is null.

For example:

// In properties

[NullAllowed]
UIImage IconFile { get; set; }

// In methods
void SetImage ([NullAllowed] UIImage image, State forState);

OverrideAttribute

Use this attribute to instruct the binding generator that the binding for this particular method should be flagged with an override keyword.

PreSnippetAttribute

You can use this attribute to inject some code to be inserted after the input parameters have been validated, but before the code calls into Objective-C.

Example:

[Export ("demo")]
[PreSnippet ("var old = ViewController;")]
void Demo ();

PrologueSnippetAttribute

You can use this attribute to inject some code to be inserted before any of the parameters are validated in the generated method.

Example:

[Export ("demo")]
[Prologue ("Trace.Entry ();")]
void Demo ();

PostGetAttribute

Instructs the binding generator to invoke the specified property from this class to fetch a value from it.

This property is typically used to refresh the cache that points to reference objects that keep the object graph referenced. Usually it shows up in code that has operations like Add/Remove. This method is used so that after elements are added or removed that the internal cache be updated to ensure that we are keeping managed references to objects that are actually in use. This is possible because the binding tool generates a backing field for all reference objects in a given binding.

Example:

[BaseType (typeof (NSObject))]
public interface NSOperation {
    [Export ("addDependency:")][PostGet ("Dependencies")]
    void AddDependency (NSOperation op);

    [Export ("removeDependency:")][PostGet ("Dependencies")]
    void RemoveDependency (NSOperation op);

    [Export ("dependencies")]
    NSOperation [] Dependencies { get; }
}

In this case, the Dependencies property will be invoked after adding or removing dependencies from the NSOperation object, ensuring that we have a graph that represents the actual loaded objects, preventing both memory leaks as well as memory corruption.

PostSnippetAttribute

You can use this attribute to inject some C# source code to be inserted after the code has invoked the underlying Objective-C method

Example:

[Export ("demo")]
[PostSnippet ("if (old != null) old.DemoComplete ();")]
void Demo ();

ProxyAttribute

This attribute is applied to return values to flag them as being proxy objects. Some Objective-C APIs return proxy objects that can not be differentiated from user bindings. The effect of this attribute is to flag the object as being a DirectBinding object. For a scenario in Xamarin.Mac, you can see the discussion on this bug.

ReleaseAttribute (Xamarin.iOS 6.0)

This can be applied to return types to indicate that the generator should call Release on the object before returning it. This is only needed when a method gives you a retained object (as opposed to an autoreleased object, which is the most common scenario)

Example:

[Export ("getAndRetainObject")]
[return: Release ()]
NSObject GetAndRetainObject ();

Additionally this attribute is propagated to the generated code, so that the Xamarin.iOS runtime knows it must retain the object upon returning to Objective-C from such a function.

SealedAttribute

Instructs the generator to flag the generated method as sealed. If this attribute is not specified, the default is to generate a virtual method (either a virtual method, an abstract method or an override depending on how other attributes are used).

StaticAttribute

When the [Static] attribute is applied to a method or property, this generates a static method or property. If this attribute is not specified, then the generator produces an instance method or property.

TransientAttribute

Use this attribute to flag properties whose values are transient, that is, objects that are created temporarily by iOS but are not long-lived. When this attribute is applied to a property, the generator does not create a backing field for this property, which means that the managed class does not keep a reference to the object.

WrapAttribute

In the design of the Xamarin.iOS/Xamarin.Mac bindings, the [Wrap] attribute is used to wrap a weakly-typed object with a strongly-typed object. This comes into play mostly with Objective-C delegate objects which are typically declared as being of type id or NSObject. The convention used by Xamarin.iOS and Xamarin.Mac is to expose those delegates or data sources as being of type NSObject and are named using the convention "Weak" + the name being exposed. An id delegate property from Objective-C would be exposed as an NSObject WeakDelegate { get; set; } property in the API contract file.

But typically the value that is assigned to this delegate is of a strong type, so we surface the strong type and apply the [Wrap] attribute, this means that users can choose to use weak types if they need some fine-control or if they need to resort to low-level tricks, or they can use the strongly-typed property for most of their work.

Example:

[BaseType (typeof (NSObject))]
interface Demo {
     [Export ("delegate"), NullAllowed]
     NSObject WeakDelegate { get; set; }

     [Wrap ("WeakDelegate")]
     DemoDelegate Delegate { get; set; }
}

[BaseType (typeof (NSObject))]
[Model][Protocol]
interface DemoDelegate {
    [Export ("doDemo")]
    void DoDemo ();
}

This is how the user would use the weakly-typed version of the Delegate:

// The weak case, user has to roll his own
class SomeObject : NSObject {
    [Export ("doDemo")]
    void CallbackForDoDemo () {}

}

var demo = new Demo ();
demo.WeakDelegate = new SomeObject ();

And this is how the user would use the strongly-typed version, notice that the user takes advantage of C#'s type system and is using the override keyword to declare its intent and does not have to manually decorate the method with [Export], since we did that work in the binding for the user:

// This is the strong case,
class MyDelegate : DemoDelegate {
   override void Demo DoDemo () {}
}

var strongDemo = new Demo ();
demo.Delegate = new MyDelegate ();

Another use of the [Wrap] attribute is to support strongly-typed version of methods. For example:

[BaseType (typeof (NSObject))]
interface XyzPanel {
    [Export ("playback:withOptions:")]
    void Playback (string fileName, [NullAllowed] NSDictionary options);

    [Wrap ("Playback (fileName, options?.Dictionary")]
    void Playback (string fileName, XyzOptions options);
}

When the [Wrap] attribute is applied on a method inside a type decorated with a [Category] attribute, you need to include This as the first argument since an extension method is being generated. For example:

[Wrap ("Write (This, image, options?.Dictionary, out error)")]
bool Write (CIImage image, CIImageRepresentationOptions options, out NSError error);

The members generated by [Wrap] are not virtual by default, if you need a virtual member you can set to true the optional isVirtual parameter.

[BaseType (typeof (NSObject))]
interface FooExplorer {
    [Export ("fooWithContentsOfURL:")]
    void FromUrl (NSUrl url);

    [Wrap ("FromUrl (NSUrl.FromString (url))", isVirtual: true)]
    void FromUrl (string url);
}

[Wrap] can also be used directly in property getters and setters. This allows to have full control on them and adjust the code as needed. For example, consider the following API definition that uses smart enums:

// Smart enum.
enum PersonRelationship {
        [Field (null)]
        None,

        [Field ("FMFather", "__Internal")]
        Father,

        [Field ("FMMother", "__Internal")]
        Mother
}

Interface definition:

// Property definition.

[Export ("presenceType")]
NSString _PresenceType { get; set; }

PersonRelationship PresenceType {
    [Wrap ("PersonRelationshipExtensions.GetValue (_PresenceType)")]
    get;
    [Wrap ("_PresenceType = value.GetConstant ()")]
    set;
}

Parameter attributes

This section describes the attributes that you can apply to the parameters in a method definition as well as the [NullAttribute] that applies to a property as a whole.

BlockCallback

This attribute is applied to parameter types in C# delegate declarations to notify the binder that the parameter in question conforms to the Objective-C block calling convention and should marshal it in this way.

This is typically used for callbacks that are defined like this in Objective-C:

typedef returnType (^SomeTypeDefinition) (int parameter1, NSString *parameter2);

See also: CCallback.

CCallback

This attribute is applied to parameter types in C# delegate declarations to notify the binder that the parameter in question conforms to the C ABI function pointer calling convention and should marshal it in this way.

This is typically used for callbacks that are defined like this in Objective-C:

typedef returnType (*SomeTypeDefinition) (int parameter1, NSString *parameter2);

See also: BlockCallback.

Params

You can use the [Params] attribute on the last array parameter of a method definition to have the generator inject a "params" in the definition. This allows the binding to easily allow for optional parameters.

For example, the following definition:

[Export ("loadFiles:")]
void LoadFiles ([Params]NSUrl [] files);

Allows the following code to be written:

foo.LoadFiles (new NSUrl (url));
foo.LoadFiles (new NSUrl (url1), new NSUrl (url2), new NSUrl (url3));

This has the added advantage that it does not require users to create an array purely for passing elements.

PlainString

You can use the [PlainString] attribute in front of string parameters to instruct the binding generator to pass the string as a C string, instead of passing the parameter as an NSString.

Most Objective-C APIs consume NSString parameters, but a handful of APIs expose a char * API for passing strings, instead of the NSString variation. Use [PlainString] in those cases.

For example, the following Objective-C declarations:

- (void) setText: (NSString *) theText;
- (void) logMessage: (char *) message;

Should be bound like this:

[Export ("setText:")]
void SetText (string theText);

[Export ("logMessage:")]
void LogMessage ([PlainString] string theText);

RetainAttribute

Instructs the generator to keep a reference to the specified parameter. The generator will provide the backing store for this field or you can specify a name (the WrapName) to store the value at. This is useful to hold a reference to a managed object that is passed as a parameter to Objective-C and when you know that Objective-C will only keep this copy of the object. For instance, an API like SetDisplay (SomeObject) would use this attribute as it is likely that the SetDisplay could only display one object at a time. If you need to keep track of more than one object (for example, for a Stack-like API) you would use the [RetainList] attribute.

Syntax:

public class RetainAttribute {
    public RetainAttribute ();
    public RetainAttribute (string wrapName);
    public string WrapName { get; }
}

TransientAttribute

This attribute is applied to parameters and is only used when transitioning from Objective-C to C#. During those transitions the various Objective-C NSObject parameters are wrapped into a managed representation of the object.

The runtime will take a reference to the native object and keep the reference until the last managed reference to the object is gone, and the GC has a chance to run.

In a few cases, it is important for the C# runtime to not keep a reference to the native object. This sometimes happens when the underlying native code has attached a special behavior to the lifecycle of the parameter. For example: the destructor for the parameter will perform some cleanup action, or dispose some precious resource.

This attribute informs the runtime that you desire the object to be disposed if possible when returning back to Objective-C from your overwritten method.

The rule is simple: if the runtime had to create a new managed representation from the native object, then at the end of the function, the retain count for the native object will be dropped, and the Handle property of the managed object will be cleared. This means that if you kept a reference to the managed object, that reference will become useless (invoking methods on it will throw an exception).

If the object passed was not created, or if there was already an outstanding managed representation of the object, the forced dispose does not take place.

Property attributes

NotImplementedAttribute

This attribute is used to support an Objective-C idiom where a property with a getter is introduced in a base class, and a mutable subclass introduces a setter.

Since C# does not support this model, the base class needs to have both the setter and the getter, and a subclass can use the OverrideAttribute.

This attribute is only used in property setters, and is used to support the mutable idiom in Objective-C.

Example:

[BaseType (typeof (NSObject))]
interface MyString {
    [Export ("initWithValue:")]
    IntPtr Constructor (string value);

    [Export ("value")]
    string Value {
        get;

    [NotImplemented ("Not available on MyString, use MyMutableString to set")]
        set;
    }
}

[BaseType (typeof (MyString))]
interface MyMutableString {
    [Export ("value")]
    [Override]
    string Value { get; set; }
}

Enum attributes

Mapping NSString constants to enum values is a easy way to create better .NET API. It:

  • allows code completion to be more useful, by showing only the correct values for the API;
  • adds type safety, you cannot use another NSString constant in a incorrect context; and
  • allows to hide some constants, making code completion show shorter API list without losing functionality.

Example:

enum NSRunLoopMode {

    [DefaultEnumValue]
    [Field ("NSDefaultRunLoopMode")]
    Default,

    [Field ("NSRunLoopCommonModes")]
    Common,

    [Field (null)]
    Other = 1000
}

From the above binding definition the generator will create the enum itself and will also create a *Extensions static type that includes two-ways conversion methods between the enum values and the NSString constants. This means the constants remains available to developers even if they are not part of the API.

Examples:

// using the NSString constant in a different API / framework / 3rd party code
CallApiRequiringAnNSString (NSRunLoopMode.Default.GetConstant ());
// converting the constants from a different API / framework / 3rd party code
var constant = CallApiReturningAnNSString ();
// back into an enum value
CallApiWithEnum (NSRunLoopModeExtensions.GetValue (constant));

DefaultEnumValueAttribute

You can decorate one enum value with this attribute. This will become the constant being returned if the enum value is not known.

From the example above:

var x = (NSRunLoopMode) 99;
Call (x.GetConstant ()); // NSDefaultRunLoopMode will be used

If no enum value is decorated then a NotSupportedException will be thrown.

ErrorDomainAttribute

Error codes are bound as an enum values. There's generally an error domain for them and it's not always easy to find which one applies (or if one even exists).

You can use this attribute to associate the error domain with the enum itself.

Example:

[Native]
[ErrorDomain ("AVKitErrorDomain")]
public enum AVKitError : nint {
    None = 0,
    Unknown = -1000,
    PictureInPictureStartFailed = -1001
}

You can then call the extension method GetDomain to get the domain constant of any error.

FieldAttribute

This is the same [Field] attribute used for constants inside type. It can also be used inside enums to map a value with a specific constant.

A null value can be used to specify which enum value should be returned if a null NSString constant is specified.

From the example above:

var constant = NSRunLoopMode.NewInWatchOS3; // will be null in watchOS 2.x
Call (NSRunLoopModeExtensions.GetValue (constant)); // will return 1000

If no null value is present then an ArgumentNullException will be thrown.

Global attributes

Global attributes are either applied using the [assembly:] attribute modifier like the [LinkWithAttribute] or can be used anywhere, like the availability attributes.

LinkWithAttribute

This is an assembly-level attribute which allows developers to specify the linking flags required to reuse a bound library without forcing the consumer of the library to manually configure the gcc_flags and extra mtouch arguments passed to a library.

Syntax:

// In properties
[Flags]
public enum LinkTarget {
    Simulator    = 1,
    ArmV6    = 2,
    ArmV7    = 4,
    Thumb    = 8,
}

[AttributeUsage(AttributeTargets.Assembly, AllowMultiple=true)]
public class LinkWithAttribute : Attribute {
    public LinkWithAttribute ();
    public LinkWithAttribute (string libraryName);
    public LinkWithAttribute (string libraryName, LinkTarget target);
    public LinkWithAttribute (string libraryName, LinkTarget target, string linkerFlags);
    public bool ForceLoad { get; set; }
    public string Frameworks { get; set; }
    public bool IsCxx { get; set;  }
    public string LibraryName { get; }
    public string LinkerFlags { get; set; }
    public LinkTarget LinkTarget { get; set; }
    public bool NeedsGccExceptionHandling { get; set; }
    public bool SmartLink { get; set; }
    public string WeakFrameworks { get; set; }
}

This attribute is applied at the assembly level, for example, this is what the CorePlot bindings use:

[assembly: LinkWith ("libCorePlot-CocoaTouch.a", LinkTarget.ArmV7 | LinkTarget.ArmV7s | LinkTarget.Simulator, Frameworks = "CoreGraphics QuartzCore", ForceLoad = true)]

When you use the [LinkWith] attribute, the specified libraryName is embedded into the resulting assembly, allowing users to ship a single DLL that contains both the unmanaged dependencies as well as the command line flags necessary to properly consume the library from Xamarin.iOS.

It's also possible to not provide a libraryName, in which case the LinkWith attribute can be used to only specify additional linker flags:

[assembly: LinkWith (LinkerFlags = "-lsqlite3")]

LinkWithAttribute constructors

These constructors allow you to specify the library to link with and embed into your resulting assembly, the supported targets that the library supports and any optional library flags that are necessary to link with the library.

Note that the LinkTarget argument is inferred by Xamarin.iOS and does not need to be set.

Examples:

// Specify additional linker:
[assembly: LinkWith (LinkerFlags = "-sqlite3")]

// Specify library name for the constructor:
[assembly: LinkWith ("libDemo.a");

// Specify library name, and link target for the constructor:
[assembly: LinkWith ("libDemo.a", LinkTarget.Thumb | LinkTarget.Simulator);

// Specify only the library name, link target and linker flags for the constructor:
[assembly: LinkWith ("libDemo.a", LinkTarget.Thumb | LinkTarget.Simulator, SmartLink = true, ForceLoad = true, IsCxx = true);

LinkWithAttribute.ForceLoad

The ForceLoad property is used to decide whether or not the -force_load link flag is used for linking the native library. For now, this should always be true.

LinkWithAttribute.Frameworks

If the library being bound has a hard requirement on any frameworks (other than Foundation and UIKit), you should set the Frameworks property to a string containing a space-delimited list of the required platform frameworks. For example, if you are binding a library that requires CoreGraphics and CoreText, you would set the Frameworks property to "CoreGraphics CoreText".

LinkWithAttribute.IsCxx

Set this property to true if the resulting executable needs to be compiled using a C++ compiler instead of the default, which is a C compiler. Use this if the library that you are binding was written in C++.

LinkWithAttribute.LibraryName

The name of the unmanaged library to bundle. This is a file with the extension ".a" and it can contain object code for multiple platforms (for example, ARM and x86 for the simulator).

Earlier versions of Xamarin.iOS checked the LinkTarget property to determine the platform your library supported, but this is now auto-detected, and the LinkTarget property is ignored.

LinkWithAttribute.LinkerFlags

The LinkerFlags string provides a way for binding authors to specify any additional linker flags needed when linking the native library into the application.

For example, if the native library requires libxml2 and zlib, you would set the LinkerFlags string to "-lxml2 -lz".

LinkWithAttribute.LinkTarget

Earlier versions of Xamarin.iOS checked the LinkTarget property to determine the platform your library supported, but this is now auto-detected, and the LinkTarget property is ignored.

LinkWithAttribute.NeedsGccExceptionHandling

Set this property to true if the library that you are linking requires the GCC Exception Handling library (gcc_eh)

The SmartLink property should be set to true to let Xamarin.iOS determine whether ForceLoad is required or not.

LinkWithAttribute.WeakFrameworks

The WeakFrameworks property works the same way as the Frameworks property, except that at link-time, the -weak_framework specifier is passed to gcc for each of the listed frameworks.

WeakFrameworks makes it possible for libraries and applications to weakly link against platform frameworks so that they can optionally use them if they are available but do not take a hard dependency on them which is useful if your library is meant to add extra features on newer versions of iOS. For more information on weak linking, see Apple's documentation on Weak Linking.

Good candidates for weak linking would be Frameworks like Accounts, CoreBluetooth, CoreImage, GLKit, NewsstandKit and Twitter since they are only available in iOS 5.

AdviceAttribute

Use this attribute to give developers a hint about other APIs that might be more convenient for them to use. For example, if you provide a strongly-typed version of an API, you could use this attribute on the weakly-typed attribute to direct the developer to the better API.

The information from this attribute is shown in the documentation and tools can be developed to give user suggestions on how to improve

RequiresSuperAttribute

This is a specialized subclass of the [Advice] attribute that can be used to hint to the developer that overriding a method requires a call to the base (overridden) method.

This corresponds to clang __attribute__((objc_requires_super))

ZeroCopyStringsAttribute

Only available in Xamarin.iOS 5.4 and newer.

This attribute instructs the generator that the binding for this specific library (if applied with [assembly:]) or type should use the fast zero-copy string marshaling. This attribute is equivalent to passing the command line option --zero-copy to the generator.

When using zero-copy for strings, the generator effectively uses the same C# string as the string that Objective-C consumes without incurring the creation of a new NSString object and avoiding copying the data from the C# strings to the Objective-C string. The only drawback of using Zero Copy strings is that you must ensure that any string property that you wrap that happens to be flagged as retain or copy has the [DisableZeroCopy] attribute set. This is require because the handle for zero-copy strings is allocated on the stack and is invalid upon the function return.

Example:

[ZeroCopyStrings]
[BaseType (typeof (NSObject))]
interface MyBinding {
    [Export ("name")]
    string Name { get; set; }

    [Export ("domain"), NullAllowed]
    string Domain { get; set; }

    [DisablZeroCopy]
    [Export ("someRetainedNSString")]
    string RetainedProperty { get; set; }
}

You can also apply the attribute at the assembly level, and it will apply to all the types of the assembly:

[assembly:ZeroCopyStrings]

Strongly-typed dictionaries

With Xamarin.iOS 8.0 we introduced support for easily creating strongly-typed classes that wrap NSDictionaries.

While it has always been possible to use the DictionaryContainer data type together with a manual API, it is now a lot simpler to do this. For more information, see Surfacing Strong Types.

StrongDictionary

When this attribute is applied to an interface, the generator will produce a class with the same name as the interface that derives from DictionaryContainer and turns each property defined in the interface into a strongly-typed getter and setter for the dictionary.

This automatically generates a class that can be instantiated from an existing NSDictionary or that has been created new.

This attribute takes one parameter, the name of the class containing the keys that are used to access the elements on the dictionary. By default each property in the interface with the attribute will lookup a member in the specified type for a name with the suffix "Key".

For example:

[StrongDictionary ("MyOptionKeys")]
interface MyOption {
    string Name { get; set; }
    nint    Age  { get; set; }
}

[Static]
interface MyOptionKeys {
    // In Objective-C this is "NSString *MYOptionNameKey;"
    [Field ("MYOptionNameKey")]
    NSString NameKey { get; }

    // In Objective-C this is "NSString *MYOptionAgeKey;"
    [Field ("MYOptionAgeKey")]
    NSString AgeKey { get; }
}

In the above case, the MyOption class will produce a string property for Name that will use the MyOptionKeys.NameKey as the key into the dictionary to retrieve a string. And will use the MyOptionKeys.AgeKey as the key into the dictionary to retrieve an NSNumber which contains an int.

If you want to use a different key, you can use the export attribute on the property, for example:

[StrongDictionary ("MyColoringKeys")]
interface MyColoringOptions {
    [Export ("TheName")]  // Override the default which would be NameKey
    string Name { get; set; }

    [Export ("TheAge")] // Override the default which would be AgeKey
    nint    Age  { get; set; }
}

[Static]
interface MyColoringKeys {
    // In Objective-C this is "NSString *MYColoringNameKey"
    [Field ("MYColoringNameKey")]
    NSString TheName { get; }

    // In Objective-C this is "NSString *MYColoringAgeKey"
    [Field ("MYColoringAgeKey")]
    NSString TheAge { get; }
}

Strong dictionary types

The following data types are supported in the StrongDictionary definition:

C# Interface Type NSDictionary Storage Type
bool Boolean stored in an NSNumber
Enumeration values integer stored in an NSNumber
int 32-bit integer stored in an NSNumber
uint 32-bit unsigned integer stored in an NSNumber
nint NSInteger stored in an NSNumber
nuint NSUInteger stored in an NSNumber
long 64-bit integer stored in an NSNumber
float 32-bit integer stored as an NSNumber
double 64-bit integer stored as an NSNumber
NSObject and subclasses NSObject
NSDictionary NSDictionary
string NSString
NSString NSString
C# Array of NSObject NSArray
C# Array of enumerations NSArray containing NSNumber values