Table of Contents Previous Next
Logo
Client-Side Slice-to-Objective-C Mapping : 18.15 Mapping for Classes
Copyright © 2003-2010 ZeroC, Inc.

18.15 Mapping for Classes

Slice classes are mapped similar to structures and exceptions. The generated class contains an instance variable and a property for each Slice data member. Consider the following class definition:
class TimeOfDay {
    short hour;         // 0  23
    short minute;       // 0  59
    short second;       // 0  59
    string format();    // Return time as hh:mm:ss
};
The Slice compiler generates the following code for this definition:
@interface EXTimeOfDay : ICEObject
{
    ICEShort hour;
    ICEShort minute;
    ICEShort second;
}

@property(nonatomic, assign) ICEShort hour;
@property(nonatomic, assign) ICEShort minute;
@property(nonatomic, assign) ICEShort second;

(id) init:(ICEShort)hour minute:(ICEShort)minute
           second:(ICEShort)second;
+(id) timeOfDay;
+(id) timeOfDay:(ICEShort)hour minute:(ICEShort)minute
                second:(ICEShort)second;
@end
There are a number of things to note about the generated code:
1. The generated class EXTimeOfDay derives from ICEObject, which is the parent of all classes. Note that ICEObject is not the same as ICEObjectPrx. In other words, you cannot pass a class where a proxy is expected and vice versa.
2. The generated class contains a property for each Slice data member.
3. The generated class provides an init method that accepts one argument for each data member, and it provides the same two convenience constructors as structures and exceptions.

18.15.1 Derivation from ICEObject

All classes ultimately derive from a common base class, ICEObject. Note that this is not the same as implementing the ICEObjectPrx protocol (which is implemented by proxies). As a result, you cannot pass a class where a proxy is expected (and vice versa) because the base types for classes and proxies are not compatible.
ICEObject defines a number of methods:
@protocol ICEObject <NSObject>
(BOOL) ice_isA:(NSString*)typeId current:(ICECurrent*)current;
(void) ice_ping:(ICECurrent*)current;
(NSString*) ice_id:(ICECurrent*)current;
(NSArray*) ice_ids:(ICECurrent*)current;
@end

@interface ICEObject NSObject <ICEObject, NSCopying>
(BOOL) ice_isA:(NSString*)typeId;
(void) ice_ping;
(NSString*) ice_id;
(NSArray*) ice_ids;
+(NSString*) ice_staticId;
(void) ice_preMarshal;
(void) ice_postUnmarshal;
(BOOL) ice_dispatch:(id<ICERequest>)request;
(id) initWithDelegate:(id)delegate;
+(id) objectWithDelegate:(id)delegate;
@end
The methods of ICEObject behave as follows:1
• ice_isA
This function returns YES if the object supports the given type ID, and NO otherwise.
• ice_ping
ice_ping provides a basic reachability test for the class. If it completes without raising an exception, the class exists and is reachable.2
• ice_ids
This function returns a string sequence representing all of the type IDs supported by this object, including ::Ice::Object.
• ice_id
This function returns the actual run-time type ID for a class. If you call ice_id via a pointer to a base instance, the returned type ID is the actual (possibly more derived) type ID of the instance.
• ice_staticId
This function returns the static type ID of a class.
• ice_preMarshal
The Ice run time invokes this function prior to marshaling the object’s state, providing the opportunity for a subclass to validate its declared data members.
• ice_postUnmarshal
The Ice run time invokes this function after unmarshaling an object’s state. A subclass typically overrides this function when it needs to perform additional initialization using the values of its declared data members.
• ice_dispatch
This function dispatches an incoming request to a servant. It is used in the implementation of dispatch interceptors (see Section 32.23).
• initWithDelegate
objectWithDelegate
These constructors enable the implementation of servants with a delegate (see page 632).

18.15.2 Data Members of Classes

By default, data members of classes are mapped exactly as for structures and exceptions: for each data member in the Slice definition, the generated class contains a corresponding property.

18.15.3 Class Constructors

Classes provide the usual init method and a parameter-less convenience constructor that perform default initialization of the class’s instance variables. In addition, if a class has data members, it provides an init method and a conve­nience constructor that accept one argument for each data member. This allows you to allocate and initialize a class instance in a single statement (instead of first having to allocate and default-initialize the instance and then assigning to its prop­erties).
For derived classes, the init method and the convenience constructor have one parameter for each of the base class’s data members, plus one parameter for each of the derived class’s data members, in base-to-derived order. For example:
class Base {
    int i;
};

class Derived extends Base {
    string s;
};
This generates:
@interface EXBase : ICEObject
// ...

@property(nonatomic, assign) ICEInt i;

(id) init:(ICEInt)i;
+(id) base;
+(id) base:(ICEInt)i;
@end

@interface EXDerived : EXBase
// ...

@property(nonatomic, retain) NSString *s;

(id) init:(ICEInt)i s:(NSString *)s;
+(id) derived;
+(id) derived:(ICEInt)i s:(NSString *)s;
@end

18.15.4 Derived Classes

Note that, in the preceding example, the derivation of the Slice definitions is preserved for the generated classes: EXBase derives from ICEObject, and EXDerived derives from EXBase. This allows you to treat and pass classes polymorphically: you can always pass an EXDerived instance where an EXBase instance is expected.

18.15.5 Passing Classes as Parameters

Classes are passed by pointer, like any other Objective‑C object. For example, here is an operation that accepts a Base as an in-parameter and returns a Derived:
Derived getDerived(Base d);
The corresponding proxy method looks as follows:
(EXDerived *) getDerived:(EXBase *)d;
To pass a null instance, you simply pass nil.

18.15.6 Operations of Classes

If you look back at the code that is generated for the EXTimeOfDay class (see page 598), you will notice that there is no indication at all that the class has a format operation. As opposed to proxies, classes do not implement any protocol that would define which operations are available. This means that you can partially implement the operations of a class. For example, you might have a Slice class with five operations that is returned from a server to a client. If the client uses only one of the five operations, the client-side code needs to implement only that one operation and can leave the remaining four operations without implemen­tation. (If the class were to implement a mandatory protocol, the client-side code would have to implement all operations in order to avoid a compiler warning.)
Of course, you must implement those operations that you actually intend to call. The mapping of operations for classes follows the server-side mapping for operations on interfaces: parameter types and labels are exactly the same. (See Section 20.5 for details.) In a nutshell, the server-side mapping is the same as the client-side mapping except that, for types that have mutable and immutable vari­ants, they map to the immutable variant where the client-side mapping uses the mutable variant, and vice versa.
For example, here is how we could implement the format operation of our TimeOfDay class:
@interface TimeOfDayI : EXTimeOfDay
@end

@implementation TimeOfDayI
(NSString *) format
{
    return [NSString stringWithFormat:@"%.2d:%.2d:%.2d",
                         self.hour, self.minute, self.second];
}
@end
By convention, the implementation of classes with operations has the same name as the Slice class with an I‑suffix. Doing this is not mandatory—you can call your implementation class anything you like. However, if you do not want to use the I‑suffix naming, we recommend that you adopt another naming convention and follow it consistently.
Note that TimeOfDayI derives from EXTimeOfDay. This is because, as we will see in a moment, the Ice run time will instantiate a TimeOfDayI instance whenever it receives a TimeOfDay instance over the wire and expects that instance to provide the properties of EXTimeOfDay.

18.15.7 Class Factories

Having created a class such as TimeOfDayI, we have an implementation and we can instantiate the TimeOfDayI class, but we cannot receive it as the return value or as an out-parameter from an operation invocation. To see why, consider the following simple interface:
interface Time {
    TimeOfDay get();
};
When a client invokes the get operation, the Ice run time must instantiate and return an instance of the TimeOfDayI class. However, unless we tell it, the Ice run time cannot magically know that we have created a TimeOfDayI class that implements a format method. To allow the Ice run time to instantiate the correct object, we must provide a factory that knows that the Slice TimeOfDay class is implemented by our TimeOfDayI class. The Ice::Communicator interface provides us with the necessary operations:
["objc:prefix:ICE"]
module Ice {
    local interface ObjectFactory {
        Object create(string type);
        void destroy();
    };

    local interface Communicator {
        void addObjectFactory(ObjectFactory factory, string id);
        ObjectFactory findObjectFactory(string id);
        // ...
    };
};
To supply the Ice run time with a factory for our TimeOfDayI class, we must implement the ObjectFactory interface:
module Ice {
    local interface ObjectFactory {
        Object create(string type);
        void destroy();
    };
};
The object factory’s create operation is called by the Ice run time when it needs to instantiate a TimeOfDay class. The factory’s destroy operation is called by the Ice run time when its communicator is destroyed. A possible implementation of our object factory is:
@interface ObjectFactory<ICEObjectFactory>
@end

@implementation ObjectFactory
(ICEObject*) create:(NSString *)type
{
    NSAssert([type isEqualToString:@"::Example::TimeOfDay"]);
    return [[TimeOfDayI alloc] init];
}
@end
The create method is passed the type ID (see Section 4.13) of the class to instantiate. For our TimeOfDay class, the type ID is "::Example::TimeOfDay". Our implementation of create checks the type ID: if it is "::Example::TimeOfDay", it instantiates and returns a TimeOfDayI object. For other type IDs, it asserts because it does not know how to instantiate other types of objects.
Note that your factory must not autorelease the returned instance. The Ice run time takes care of the necessary memory management activities on your behalf.
Given a factory implementation, such as our ObjectFactory, we must inform the Ice run time of the existence of the factory:
id<ICECommunicator> ice = ...;
ObjectFactory *factory =
    [[[ObjectFactory alloc] init] autorelease];
[ic addObjectFactory:factory sliceId:@"::Example::TimeOfDay"];
Now, whenever the Ice run time needs to instantiate a class with the type ID "::Example::TimeOfDay", it calls the create method of the registered ObjectFactory instance.
The destroy operation of the object factory is invoked by the Ice run time when the communicator is destroyed. This gives you a chance to clean up any resources that may be used by your factory. Do not call destroy on the factory while it is registered with the communicator—if you do, the Ice run time has no idea that this has happened and, depending on what your destroy implementa­tion is doing, may cause undefined behavior when the Ice run time tries to next use the factory.
The run time guarantees that destroy will be the last call made on the factory, that is, create will not be called concurrently with destroy, and create will not be called once destroy has been called. However, the Ice run time may make concurrent calls to create.
Note that you cannot register a factory for the same type ID twice: if you call addObjectFactory with a type ID for which a factory is registered, the Ice run time throws an AlreadyRegisteredException.
Finally, keep in mind that if a class has only data members, but no operations, you need not (but can) create and register an object factory to transmit instances of such a class. Only if a class has operations do you have to define and register an object factory.

Using a Category to Implement Operations

An alternative to registering a class factory is to use an Objective-C category to implement operations. For example, we could have implemented our format method using a category instead:
@interface EXTimeOfDay (TimeOfDayI)
@end

@implementation EXTimeOfDay (TimeOfDayI)
(NSString *) format
{
    return [NSString stringWithFormat:@"%.2d:%.2d:%.2d",
                         self.hour, self.minute, self.second];
}
@end
In this case, there is no need to derive from the generated EXTimeOfDay class because we provide the format implementation as a category. There is also no need to register a class factory: the Ice run time instantiates an EXTimeOfDay instance when a TimeOfDay instance arrives over the wire, and the format method is found at run time when it is actually called.
This is a viable alternative approach to implement class operations. However, keep in mind that, if the operation implementation requires use of instance vari­ables that are not defined as part of the Slice definitions of a class, you cannot use this approach because Objective‑C categories do not permit you to add instance variables to a class.

18.15.8 Copying of Classes

Classes implement NSCopying. The behavior is the same as for structures: instance variables of value type are copied by assignment, instance variables of pointer type are copied by calling retain, that is, the copy is shallow. To illustrate this, consider the following class definition:
class Node {
    int i;
    string s;
    Node next;
};
We can initialize two instances of type EXNode as follows:
NSString lastString = [NSString stringWithString:@"last"];
EXNode *last = [EXNode node:99 s:lastString next:nil];

NSString firstString = [NSString stringWithString:@"first"];
EXNode *first = [EXNode node:1 s:firstString next:last];
This creates the situation shown in Figure 18.2.
Figure 18.2. Two instances of type EXNode.
Now we create a copy of the first node by calling copy:
EXNode *copy = [[first copy] autorelease];
This creates the situation shown in Figure 18.3.
Figure 18.3. EXNode instances after calling copy on first.
As you can see, the first node is copied, but the last node (pointed at by the next instance variable of the first node) is not copied; instead, first and copy now both have their next instance variable point at the same last node, and both point at the same string.

Cyclic References

One thing to be aware about are cyclic references among classes. As an example, we can easily create a cycle by executing the following statements:
EXNode *first = [EXNode node];
ExNode *last = [EXNode node];
first.next = last;
last.next = first;
This makes the next instance variable of the two classes point at each other, creating the cycle shown in Figure 18.4.
Figure 18.4. Two nodes with cyclic references.
There is no problem with sending this class graph as a parameter. For example, you could pass either first or last as a parameter to an operation and, in the server, the Ice run time will faithfully rebuild the corresponding graph, preserving the cycle. However, if a server returns such a graph from an operation invocation as the return value or as an out-parameter, all class instances that are part of a cycle are leaked. The same is true on the client side: if you receive such a graph from an operation invocation and do not explicitly break the cycle, you will leak all instances that form part of the cycle.
Because it is difficult to break cycles manually (and, on the server side, for return values and out-parameters, it is impossible to break them), we recommend that you avoid cyclic references among classes.3

1
The methods are split between the ICEObject protocol and class because classes can be servants.

2
ice_ping is normally only invoked on the proxy for a class that might be remote because a class instance that is local (in the caller’s address space) can always be reached.

3
A future version of the Objective‑C run time may provide a garbage collector similar to the one used by Ice for C++.


Table of Contents Previous Next
Logo