What does the unavailable attribute in Objective C do?
__attribute__((unavailable("message")))
Is there any online reference to this and other attributes in Clang?
The unavailable attribute marks a function declaration so that you can generate an error message if someone tries to use it. It's essentially the same as the deprecated attribute, except that trying to use a deprecated function just causes a warning, but using an unavailable one causes an error. Documentation at: http://clang.llvm.org/docs/LanguageExtensions.html
Here's a simple use case example. First the code:
void badFunction(void) __attribute__((unavailable("Don't use badFunction, it won't work.")));
int main(void)
{
badFunction();
return 0;
}
And then building it:
$ make example
cc example.c -o example
example.c:5:5: error: 'badFunction' is unavailable: Don't use badFunction, it
won't work.
badFunction();
^
example.c:1:6: note: function has been explicitly marked unavailable here
void badFunction(void) __attribute__((unavailable("Don't use...
^
1 error generated.
make: *** [example] Error 1
Without getting into a discussion of the pros and cons of singleton objects, attribute((unavailable("message"))) is handy for preventing singletons from being instantiated outside of the standard "sharedInstance" method.
For instance, in the header file of your singleton manager object the following lines will prevent the use of alloc, init, new or copy.
// clue for improper use (produces compile time error)
+ (instancetype)alloc __attribute__((unavailable("alloc not available, call sharedInstance instead")));
- (instancetype)init __attribute__((unavailable("init not available, call sharedInstance instead")));
+ (instancetype)new __attribute__((unavailable("new not available, call sharedInstance instead")));
- (instancetype)copy __attribute__((unavailable("copy not available, call sharedInstance instead")));
In order to instantiate your singleton you will need to roll your own custom initialization routine. Something along the lines of:
#interface singletonManager ()
-(instancetype)initUniqueInstance;
#end
#implementation singletonManager
+ (instancetype)sharedInstance
{
static id instance = nil;
static dispatch_once_t onceToken = 0;
dispatch_once(&once_token,^
{
instance = [[super alloc] initUniqueInstance];
});
return instance;
}
- (instancetype)initUniqueInstance
{
if (( self = [super init] ))
{
//regular initialization stuff
}
return self;
}
#end
Related
I am messing around with method swizzling and would like to call the original function after performing a method_exchangeImplementations. I have two projects I have setup for this.
The first project is the main project for the application. This project includes all of the logic for the application. Notice that originalMethodName is called when the view loads.
#implementation ViewController
- (void)originalMethodName
{
NSLog(#"REAL %s", __func__);
}
- (void)viewDidLoad {
[super viewDidLoad];
NSLog(#"REAL %s", __func__);
[self originalMethodName];
}
#end
The second project includes only the code for swizzling. I have a method swizzle_originalMethodName which includes the code I want to inject into the main application with the originalMethodName function is called.
#implementation swizzle_ViewController
- (void)swizzle_originalMethodName
{
NSLog(#"FAKE %s", __func__);
}
__attribute__((constructor)) static void initializer(void)
{
NSLog(#"FAKE %s", __func__);
Class c1 = objc_getClass("ViewController");
Class c2 = [swizzle_ViewController class];
Method m1 = class_getInstanceMethod(c1, #selector(originalMethodName));
Method m2 = class_getInstanceMethod(c2, #selector(swizzle_originalMethodName));
method_exchangeImplementations(m1, m2);
}
#end
The swizzle is working just fine (as seen in the output below), but now I want to be able to call originalMethodName from the swizzle_originalMethodName
2016-08-17 14:18:51.765 testMacOS[7295:1297055] FAKE initializer
2016-08-17 14:18:51.822 testMacOS[7295:1297055] REAL -[ViewController viewDidLoad]
2016-08-17 14:18:51.822 testMacOS[7295:1297055] FAKE -[swizzle_ViewController swizzle_originalMethodName]
I have tried to use NSInvocation but am not having any luck. Any ideas what I am doing wrong?
Class c1 = objc_getClass("ViewController");
Method m1 = class_getInstanceMethod(c1, #selector(originalMethodName));
NSMethodSignature *methodSignature = [NSMethodSignature signatureWithObjCTypes:method_getTypeEncoding( m1)];
NSInvocation *originalInvocation = [NSInvocation invocationWithMethodSignature:methodSignature];
[originalInvocation invoke];
If you are swizzling within a class hierarchy, e.g. you have a subclass which swizzles one of its ancestors methods with one of its own, then you simply have the swizzled-in method apparently call itself – that call will actually call the swizzled-out method as the methods have been swapped. In your case you would have:
- (void)swizzle_originalMethodName
{
NSLog(#"FAKE %s", __func__);
[self swizzle_originalMethodName]; // call original
}
This does not work in your case as you are cross-class swizzling, so self doesn't reference the class with the swizzled-out method. And you don't have an instance of the swizzling class you can call the swizzled-out method on...
Here is one easy way to fix this, what your swizzle-in method needs to be able to do is call the original implementation, and you can get that when you setup the swizzling.
In Objective-C a method is implemented by a function whose first two arguments are the object reference the method is being called on and the selector and the remaining arguments are those of the method. For example the NSString method:
- (NSRange)rangeOfString:(NSString *)aString
is implemented by a function something like:
NSRange rangeOfStringImp(NSString *self, SEL cmd, NSString *aString)
You can obtain a function pointer to this implementation function using method_getImplementation.
To your code, first in your swizzle_ViewController declare a type for the implementation function of the method you are swizzling, and a global to store the function pointer:
typedef void (*OriginalImpType)(id self, SEL selector);
static OriginalImpType originalImp;
Now in your initializer method you need to save the method implementation, you can do this by adding the line shown:
Method m1 = class_getInstanceMethod(c1, #selector(originalMethodName));
originalImp = (OriginalImpType)method_getImplementation(m1); // save the IMP of originalMethodName
Finally have your swizzled-in method call the saved implementation:
- (void)swizzle_originalMethodName
{
NSLog(#"FAKE %s", __func__);
originalImp(self, #selector(originalMethodName)); // call the original IMP with the correct self & selector
}
Optional: The above works correctly, however it does a little more than is required – the method implementations are both exchanged and one is stored in a global variable, all you really need to do is save the original implementation of m1 and then set its implementation to that of m2. You can address this by replacing the call to method_exchangeImplementations with:
method_setImplementation(m1, method_getImplementation(m2));
It is a little more typing, but somewhat clearer as to what actually needs to be done.
HTH
There is a slightly easier option to call the original implementation that doesn't require you to store the method implementation directly. When you exchange implementations of the methods, the original implementation will be stored in the swizzler class. You can fetch the swizzled out implementation using the class_getMethodImplementation function. Here is a playground sample:
import Cocoa
let fooSelector = Selector("fooWithArg:")
let swizzledFooSelector = Selector("swizzled_fooWithArg:")
class A: NSObject {
#objc dynamic func foo(arg: String) {
print("Foo \(arg) in A")
}
}
class B: NSObject {
private typealias FooFunc = #convention(c) (AnyObject, Selector, String) -> Void
#objc func swizzled_foo(arg: String) {
print("Swizzled_foo \(arg) in B")
unsafeBitCast(
class_getMethodImplementation(B.self, swizzledFooSelector),
to: FooFunc.self
)(self, fooSelector, arg)
}
}
method_exchangeImplementations(
class_getInstanceMethod(A.self, fooSelector)!,
class_getInstanceMethod(B.self, swizzledFooSelector)!
)
A().foo(arg: "bar")
I have the following method:
- (FDModel *)_modelForClass: (Class)modelClass
withIdentifier: (NSString *)identifier
which should take in a Class and a identifier, create an instance of modelClass, assign the identifier and do some other work based on the fact that it assumed modelClass is a subclass of FDModel.
I can put in a check that raises some error or exception if [modelClass isSubclassOfClass: [FDModel class]] == NO but I was trying to see if there was a way to enforce this at compile time.
EDIT: I understand that some people see this as a obvious factory method but the modelClass parameter is actually passed in by the user of my library through a delegate callback - (Class<FDModel>)modelClassForDictionary: (NSDictionary *)dictionary;. This question was more aimed at making the user of my library return a Class that has a specific subclass.
I would consider the plain answer to your question being no; there is no way of checking if a class passed as a parameter is of a certain kind.
But I'd like to argue that the essence of your question primarily points to a design issue, i.e. can't your instance-generating method be expressed as a factory method? Like so:
#interface FDModel
+ (instancetype)modelWithIdentifier:(NSString *)identifier;
#end
In the above case you would simply do:
[FDModel modelWithIdentifier:anIdentifier];
The actual class returned (and the initialisation logic) being specified by the factory method implementation through subclassing of the FDModel class:
#implementation FDModelSubclass
+ (instancetype)modelWithIdentifier:(NSString *)identifier
{
FDModel *model = [super modelWithIdentifier:identifier];
if (model)
{
// do additional init stuff
}
return model;
}
#end
Nothing to check, no chance to go wrong.
After some research I don't think you can do it at compile time - you have to do it at runtime as you expected.
BOOL classConformsToProtocol = [class conformsToProtocol:#protocol(OKAProtocol)];
OR
BOOL classConformsToProtocol = [self class:[OKAClass class] conformsToProtocol:#"OKAProtocol"];
------
- (BOOL)class:(Class)class conformsToProtocol:(NSString *)protocol;
{
return [class conformsToProtocol:NSProtocolFromString(protocol)];
}
In Cocoa and Objective C the favorite method for managing error seems to be using an NSError * object, to construct an error object however, we need to call the following method
+ (id)errorWithDomain:(NSString *)domain code:(NSInteger)code userInfo:(NSDictionary *)dict
My question is, what are some of the best practices for managing error domain, error code definitions and user info dictionary across the entire application so that error code, domain and user info dict always stays consistent?
If you have a hefty amount of error construction, your life could be much simpler by using a class. I actually use C++ for this so the calls a program does not need may be removed (unlike objc), but you can use C, ObjC, or C++ for this:
MONErrorDomain.h
// you won't normally need an instance here
#interface MONErrorDomain : NSObject
+ (NSString *)domain; // << required override
- (NSString *)domain; // << returns [[self class] domain]
// example convenience methods:
// uses [self domain]
+ (NSError *)errorWithErrorCode:(NSInteger)errorCode; // << user info would be nil
+ (NSError *)errorWithErrorCode:(NSInteger)errorCode userInfo:(NSDictionary *)userInfo;
#end
MONKoalaError.h
#interface MONKoalaError : MONErrorDomain
+ (NSError *)outOfEucalyptus;
#end
extern NSString * const MONKoalaErrorDomain;
typedef enum MONKoalaErrorCode {
MONKoalaErrorCode_Undefined = 0,
MONKoalaErrorCode_OutOfEucalyptus
} MONKoalaErrorCode;
MONKoalaError.m
// apple recommends we use reverse domains
NSString * const MONKoalaErrorDomain = #"com.mon.koala-library.MONKoalaErrorDomain";
#implementation MONKoalaError
+ (NSString *)domain
{
return MONKoalaErrorDomain;
}
+ (NSError *)outOfEucalyptus
{
NSDictionary * info = …;
return [self errorWithErrorCode:MONKoalaErrorCode_OutOfEucalyptus userInfo:info];
}
#end
Then the error creation is all in one place for each domain, and the clients can easily pick their errors without actually building them manually:
if (outError) {
*outError = [MONKoalaError outOfEucalyptus];
}
and error handling takes the form:
if ([e.domain isEqualToString:MONKoalaErrorDomain]) {
switch (e.code) {
case MONKoalaErrorCode_OutOfEucalyptus : {
self.needsEucalyptus = true;
…
One common way is to define some appropriate constants in a header file, and then include that header file wherever needed. It's a pretty simple approach, and looks like:
const NSString * kMyAppErrorDomain = #"com.example.myapp";
const NSInteger kMyAppSomeError = 2;
// key into user info dictionary
const NSString * kMyAppProblemKey = #"MyAppProblemKey";
I've also seen some applications which create convenience methods for creating these, either as a category on NSError or as a separate utility class or set of functions. It's also entirely reasonable to subclass NSError, for example to customize the localized description.
If you have not already seen it, Apple has released the Error Handling Programming Guide which discusses how these should be used in Cocoa.
The following code implements an NSProxy subclass which forwards methods to an NSNumber instance.
However when calling [nsproxy floatValue] I get 0.0 under GCC 4.2.
Under LLVM-Clang I get the correct answer 42.0.
Any idea what is going on?
(by the way this is running under Garbage Collection)
-(id) init;
{
_result = [NSNumber numberWithFloat:42.0];
return self;
}
- (NSMethodSignature *)methodSignatureForSelector:(SEL)aSelector
{
return [[_result class] instanceMethodSignatureForSelector:aSelector];
}
- (void)forwardInvocation:(NSInvocation *)anInvocation
{
[anInvocation setTarget:_result];
[anInvocation invoke];
return;
}
Your proxy class doesn't declare the message signature for -floatValue. Since it returns a floating point number, that can be a problem. Because you haven't declared it anywhere and presumably aren't casting your proxy object to its' represented class the compiler has to guess at the method signature. In that case the GCC compiler guesses that the message will return an id pointer.
In Objective-C, depending on the signature of the message and the architecture of the machine, different functions get used for processing the message and its' return values. On x86 machines messages that return a floating point value are called through objc_msgSend_fpret while functions that return void and id use objc_msgSend.
Since the GCC compiler is assuming that the return value is an id it uses the latter function and incorrectly handles the result. That Clang is able to handle this properly is interesting but I would hesitate to rely on that behavior. It would be better to declare a category on your proxy for any methods that you'll be forwarding. That also has the benefit of removing the warning that was being generated for the line of code calling the floatValue method.
#interface Foo (ProxyMethods)
- (float)floatValue;
#end
Any init method should call [super init]; to prevent unexpected behaviour. As so:
- (id)init {
self = [super init];
if (self) {
// Your init code
}
return self;
}
I have a method in an objective-C class. It has 2 callback functions written in C. The class pointer i.e. self is passed to these functions as void *. In the C functions I create a pointer of type class and assign the void * parameter.
The first callback function executes successfully. But the void * pointer becomes nil in the 2nd callback function. Note that I haven't tweaked pointer in the first callback but still I get nil in 2nd callback.
Any ideas what might be going wrong?
For example:
kr = IOServiceAddMatchingNotification(gNotifyPort, kIOFirstMatchNotification,
matchingDict, RawDeviceAdded, NULL,
&gRawAddedIter);
RawDeviceAdded(NULL, gRawAddedIter, self);
This works fine. But below function receives self as nil.
kr = IOServiceAddMatchingNotification(gNotifyPort, kIOFirstMatchNotification,
matchingDict, BulkTestDeviceAdded, NULL,
&gBulkTestAddedIter);
BulkTestDeviceAdded(NULL, gBulkTestAddedIter, self);
Are your problems specifically with the IOKit callback routines? The problem with the specific example you gave is that the IOServiceMatchingCallback takes only 2 parameters, not 3. You need your RawDeviceAdded() and BulkTestDeviceAdded() callback functions to match the IOServiceMatchingCallback prototype and to accept self as the first parameter (refCon), not the 3rd. Also, you need to pass in self as the second-to-last parameter of IOServiceAddMatchingNotification() to get it passed back to you by the callback.
A common method for handling C callbacks in Objective-C code is just to have a static function that forwards the callback to your instance. So, your example callback code would look like this:
static RawDeviceAdded(void* refcon, io_iterator_t iterator)
{
[(MyClass*)refcon rawDeviceAdded:iterator];
}
#implementation MyClass
- (void)setupCallbacks
{
// ... all preceding setup snipped
kr = IOServiceAddMatchingNotification(gNotifyPort,kIOFirstMatchNotification, matchingDict,RawDeviceAdded,(void*)self,&gRawAddedIter );
// call the callback method once to 'arm' the iterator
[self rawDeviceAdded:gRawAddedIterator];
}
- (void)rawDeviceAdded:(io_iterator_t)iterator
{
// take care of the iterator here, making sure to complete iteration to re-arm it
}
#end
Generally, callbacks in Objective-C are handled by passing a delegate object and a selector to perform on that delegate. For example, this method will call a method on its delegate after logging a message, passing both itself and the message that was logged.
- (void)logMessage:(NSString *)message
delegate:(id)delegate
didLogSelector:(SEL)didLogSelector
{
NSLog(#"%#", message);
if (delegate && didLogSelector && [delegate respondsToSelector:didLogSelector]) {
(void) [delegate performSelector:didLogSelector
withObject:self
withObject:message];
}
}
You might call it in code like this:
- (void)sayHello
{
[logger logMessage:#"Hello, world"
delegate:self
didLogSelector:#selector(messageLogger:didLogMessage:)];
}
- (void)messageLogger:(id)logger
didLogMessage:(NSString *)message
{
NSLog(#"Message logger %# logged message '%#'", logger, message);
}
You can also use objc_msgSend() directly instead, though you need to understand the Objective-C runtime enough to choose which variant to use and how to construct the prototype and function pointer through which to call it. (It's the mechanism by which message sends are actually implemented in Objective-C — what the compiler normally generates calls to in order to represent [] expressions.)
This is what Objective-C's selector is for:
http://developer.apple.com/iphone/library/documentation/Cocoa/Reference/NSInvocationOperation_Class
The API isn't very intuitive, but its fine once you understand it
You might need to do some refactoring as well, now there might be a better way, but when I had this problem my solution was to refactor and use InvoationOperation.