I would like to pass a callback block from my objective-C++ code to a C++ object. It's very straightforward since I'm able to assign a block to a std::function. In my mini example everything worked fine, but I'm still uncertain if it's safe to do that.
#import "ViewController.h"
#include <functional>
std::function<void(void)> f;
#interface T : NSObject
#property (strong) NSString* value;
#end
#implementation T
- (void)dealloc {
NSLog(#"dealloc");
}
#end
#implementation ViewController
- (IBAction)storeBlock:(UIButton *)sender {
T* t = [[T alloc] init];
t.value = #"the captured obj";
f = ^void(void) { NSLog(#"This is the block with %#", t.value); };
}
- (IBAction)useBlock:(UIButton *)sender {
f();
}
- (IBAction)releaseBlock:(UIButton *)sender {
f = nullptr;
}
#end
I've learned that the blocks are stored on the stack and I have to copy it to the heap if I want to use it further than the scope of the block's creation (which I haven't done in my example explicitly). I'm also uncertain how ARC can handle this situation. Is it 'officialy' allowed to do this? I'm using Xcode 9.2.
Objective-C blocks become invalid when execution leaves the scope where the block is defined. To avoid this, the copy method must be used on the block, that's all.
Blocks are treated like normals objects despite the fact, that they start their lifetime on the stack. Like you said, in objc-cpp we can easily store a block in a std::function :-)
So what happens in your example?
You declared the std::function "f" as a global stack object. When the block is assigned to it, the reference count from the block object is 1. Your "T"-object has in the storeBlock function before leaving two references on it. One by "t" and one through the capturing mechanism from the block. Leaving the scope, only the blocks reference remains.
Then "f" is set to nullptr: There is no reference to the block object -> the block gets dealloced and with it the "T"-object, because it points a reference to it neither.
Related
What I have read from the apple document retain will increase the retain count by 1 , and release will decrease by 1. This is very much clear to me.
But In the case of copy and retain i am a bit confused.
Let me explain with the code i am trying.
property ---
#property(nonatomic, retain) NSMutableString *a;
#property(nonatomic, copy) NSMutableString *b;
#synthesize a = _a ,b = _b
a=[[NSMutableString alloc]initWithString:#"Hello Ankit"];
NSLog(#"a memory location A - %p", &a );
b=[[NSMutableString alloc]initWithString:#"Hello Nigam"];
NSLog(#"a memory location B- %p", &b );
c= [[NSMutableString alloc]initWithString:#"Ankit Nigam"];
NSLog(#"a memory location C %p",&c);
NSMutableString *temp =[[NSMutableString alloc]initWithString:#"hey"];
NSLog(#"temp = %# %p",temp,&temp);
self.b = temp;
NSLog(#"B is now %# %p",self.b,&b);
self.a = temp;
NSLog(#"A is now %# %p",self.a,&a);
And i get the output as -- - -
2012-05-10 03:24:34.756 retainCountTest[2655:f803] a memory location A - 0x6d314fc
2012-05-10 03:24:34.757 retainCountTest[2655:f803] a memory location B- 0x6d31500
2012-05-10 03:24:34.764 retainCountTest[2655:f803] a memory location C 0x6d31504
2012-05-10 03:24:34.764 retainCountTest[2655:f803] temp = hey 0xbfffdd04
2012-05-10 03:24:34.764 retainCountTest[2655:f803] B is now hey 0x6d31500
2012-05-10 03:24:34.765 retainCountTest[2655:f803] A is now hey 0x6d314fc
But as per I understand from the Doc the retain object must have the same memory address , where as copy object will create a new object with different memory location.
Again when i change the logs to ---
self.b = temp;
NSLog(#"B is now %# %p",self.b,&_b);
self.a = temp;
NSLog(#"A is now %# %p",self.a,&_a);
It return me a complete different memory location for both the object.
2012-05-10 03:28:49.905 retainCountTest[2688:f803] a memory location A - 0x6d4a4ac
2012-05-10 03:28:49.906 retainCountTest[2688:f803] a memory location B- 0x6d4a4b0
2012-05-10 03:28:49.907 retainCountTest[2688:f803] a memory location C 0x6d4a4b4
2012-05-10 03:28:49.907 retainCountTest[2688:f803] temp = hey 0xbfffdd04
2012-05-10 03:28:49.908 retainCountTest[2688:f803] B is now hey 0x6d4a4c0
2012-05-10 03:28:49.908 retainCountTest[2688:f803] a is now hey 0x6d4a4bc
Can any help me to understand the complete concept of these retain and copy. Also why I am getting these unexpected results.
Thanks a lot.
A property is just a declaration that allows for setters, getters, and dot-syntax accessors (interface variable hiding).
It does absolutely nothing on its own but allow you to use -[myInstance myProperty] to get the variable or use -[myInstance setMyProperty:] to set it (yes, the method name is auto-assigned to -setProperty: and -property).
When declaring a property, you have three categories - thread locking, access control, and memory management. You can only pick one of the modifiers for each category and if you do not pick one, it's auto-assigned to one automatically.
#property (<thread locking>, <access control>, <memory management>) id property;
The first category can either be atomic or nonatomic. The atomic modifier forces an #synchronized(myInstance) lock on the variable, to allow thread safety. The nonatomic does not use a synchronized-block, and is NOT thread safe. If you do not use either, it is automatically set to atomic.
The second category can either be readonly or readwrite. The readwrite modifier allows the property to be modified as well, and allows auto-generation of the -setProperty: method. When the readonly modifier is used, you cannot use the -setProperty: method. You must use the internal variable from within the object to set the variable directly.
The third category can either be assign, retain, and copy. The assign modifier means the internal object pointer is set to the pointer passed to the -setProperty: message. The retain modifier assigns the passed pointer and passes a -retain to the object.
The copy modifier does a straight-up clone of the object- a new pointer to a new object at a new address in the memory. This sets the internal object pointer to the copy of the passed object, by calling -copy on the passed object. The default modifier is assign, and the compiler will warn you if you do not set the memory management category modifier on an object - because an assign modifier on an object is frowned upon (unless explicitly declared).
For an example on -copy, look at this:
- (void)setProperty:(GXMyObject *)property {
// This points to the original passed object.
GXMyObject *original = property;
// This points to a copy of the passed object.
CGMyObject *copied = [property copy];
// This points to yet another copy of the passed object-
// Independent of the other copies and original.
_property = [property copy];
// The anotherProperty is now different on this copy
// than on the original and the other copies.
_property.anotherProperty = 4;
// This will prove that they are all individual objects.
NSLog(#"%p, %p, %p", original, copied, _property);
}
There is an optional method name declaration modifier and is used like so: getter = myCustomPropertyGetter and setter = myCustomPropertySetter: (The colon : at the end of the setter method name is required because it denotes that an argument must be passed).
The second half of this is the property synthesizer or dynamizer. Once a property is declared (for example, myView) like so:
#property (nonatomic, retain) NSView *myView;
You would either: define the setter and getter yourself; #synthesize the setter and getter; #dynamic the property to say that it exists in a category or the main class, or may be added at runtime (not a fun idea, mind you, and could cause bad runtime exceptions).
The first example, writing the methods yourself would look like this:
// In Apple's LLVM 3.1 Compiler, instance variables can be added
// within {} below the #implementation as well as the #interface,
// and in private categories (#interface GXMyClass ()) like before.
#implementation GXMyClass {
// The internal object pointer is prefixed with an _ to avoid name confusions.
NSView *_myView;
}
- (NSView *)myView {
return _myView;
}
- (void)setMyView:(NSView *)myView {
_myView = [myView retain];
}
#end
The second example would be to auto-synthesize it using the #synthesize directive:
#implementation GXMyClass
// In the new Apple LLVM 3.1 Clang compiler, the = operator when used
// next to the #synthesize directive declares an internal private
// variable and automatically sets to that variable.
#synthesize myView = _myView;
// The internal variable name is now myOtherView, because we did not use the
// = operator to assign an internal variable name to the property.
#synthesize myOtherView;
#end
Under the last example, perhaps the most confusing, because it requires the use of the #dynamic directive, you require something of a category or a runtime method addition:
#interface GXMyClass (InternalMethods)
#end
#implementation GXMyClass
// The = assignment operator does not work here.
#dynamic myView;
#end
#implementation GXMyClass (InternalMethods)
- (NSView *)myView {
return [self methodThatReturnsAnNSView];
}
- (void)setMyView:(NSView *)myView {
[self methodThatAcceptsAnNSViewArgument:myView];
}
#end
The #property declaration requires one of the three above declarations to be present- it does not do anything on its own. What it DOES allow, however is dot-syntax accessors (Java-like accessors for setting and getting properties).
For example, #property (copy) NSString *myName; could be accessed using -[myObject myName] and set using -[myObject setMyName:].
Now it can be set using myObjectInstance.myName = #"Bob"; and gotten using myObjectInstance.myName. Utilizing all the above concepts, one could create an object such as this one:
// The GXBufferQueue is a queue which buffers all requests, till they are read
// asynchronously later. The backing store is an NSMutableArray to which all
// buffer writes are appended to, and from which the first object is pulled and
// returned when the buffer is read to.
#interface GXBufferQueue
#property (nonatomic, readwrite, copy, setter = write:, getter = read) id buffer;
+ (GXBufferQueue *)queue;
#end
#implementation GXBufferQueue {
// This queue is an internal array and is 'tacked on' to the #implementation
// so no others can see it, and it can be marked #private so subclasses cannot
// use it. It is also good code practice to have #interfaces composed of only
// #properties, setters, and getters, rather than expose internal variables.
NSMutableArray *_internalQueue;
}
+ (GXBufferQueue *)queue {
return [[[GXBufferQueue alloc] init] autorelease];
}
- (id)init {
if((self = [super init])) {
_internalQueue = [[NSMutableArray alloc] init];
}
}
- (void)write:(id)buffer {
[_internalQueue addObject:buffer];
}
- (id)read {
if(!(_internalQueue.count > 0)) return nil;
id buffer = [_internalQueue objectAtIndex:0];
[_internalQueue removeObjectAtIndex:0];
return buffer;
}
#end
Note: This code was in no way tested.
Now that you have a GXBufferQueue, all the following works:
GXBufferQueue *queue = [GXBufferQueue queue];
// Option One: using the traditional message syntax:
[queue write:#"This will be now added to the buffer."];
NSLog(#"Now the string written to the queue will be read \
and removed from the queue, like a stack pop. ", [queue read]);
// Option Two: using the new dot-syntax accessors:
queue.buffer = #"As clunky as this looks, it works the same as above.";
NSLog(#"These lines work just the same as the ones above: ", queue.buffer);
As you can see, there's a lot possible with properties and a lot more that can be done with them than just variable declaration. If there are any questions or anything the community would like me to add/rectify to/in the post, please leave a comment! :D
Yes, retainCount is useless-- don't call it-- and assuming that copy will always return a new instance is incorrect.
If an immutable object is sent copy, it will typically be implemented as:
- copy { return [self retain]; }
I.e. there is no need to make an actual copy since the contents can't change.
Of course, since you are mucking about with static, compiled into the binary, strings, the implementation is likely closer to:
- retain { return self; }
- copy { return self; }
- (void)release { ; }
- (void)dealloc { ; }
Maybe -- all of the above are implementation details that may change at any time. The bottom line, though, is that all of the above fulfills the contract of retain/release/copy/etc...
I have an interface like this:
#interface MacCalculatorAppDelegate:NSObject
<UIApplicationDelegate> {
// ...
UIButton *operatorPressed;
NSString *waitingOperation;
}
And I am initializing waitingOperation variable in my implementation like this:
- (id)init {
if (self = [super init]) {
waitingOperation = #"not set";
}
return self;
}
And I want to reinitialize this variable in a function. This is calculator program and when user clicks on operators button the following function will be invoked:
- (IBAction)operatorPressed:(UIButton *)sender {
if([#"+" isEqual:operand]) {
waitingOperation = #"+";
}
}
But after the check in if statement, my program won't do anything and this happens when I am trying to reinitialize waitingOperation variable.
I am new to objective-c, please help me understand what's wrong here.
Thanks in advance.
There are several things to note here.
waitingOperation=#"not set";
As it stands, this will eventually crash your program. Objective C string literals are autoreleased instances of NSString, which means unless you assign it to a retained property or retain it manually, the memory will be deallocated, leaving a dangling pointer.
-(IBAction) operatorPressed:(UIButton *)sender {
Have you verified that this method is actually being called? You have to assign the IBAction in Interface Builder. Step through it in the debugger or use NSLog to verify that this method is being called.
if([#"+" isEqual:operand])
Where is operand coming from?
waitingOperation=#"+";
Same problem as above, this will get deallocated behind the scenes, leaving you with a dangling pointer.
Also, note that if you know that both variables are NSStrings, using isEqualToString: is faster than using isEqual:.
Finally, this stuff shouldn't be part of your app delegate. This is view controller logic.
If your operand is also a string, then check for isEqualToString instead of isEqual
i have some trouble writing a method in Objective-C to make an object nil. Here is some example :
#interface testA : NSObject
{
NSString *a;
}
#property (nonatomic, retain) NSString *a;
+(testA*)initWithA:(NSString *)aString;
-(void)displayA;
-(void)nillify;
#end
#implementation testA
#synthesize a;
+(testA*)initWithA:(NSString *)aString{
testA *tst=[[testA alloc] init];
tst.a=aString;
return [tst autorelease];
}
-(void)displayA{
NSLog(#"%#",self.a);
}
-(void)nillify{
self=nil;
}
- (void)dealloc {
[a release];
[super dealloc];
}
#end
int main(int argc, char **argv){
NSAutoreleasePool *pool = [[NSAutoreleasePool alloc] init];
testA *test=[testA initWithA:#"some test"];
[test displayA];
test=nil;
//[test nillify];
NSLog(#"after setting to nil");
[test displayA];
[pool release];
return 0;
}
Apparently , when I set test object to nil and then call some method on it nothing happens , but if i call nillify instead of directly setting it to nil , displayA method works normally like test object is still there. Is there a workaround for nillify method to function properly ?
Your help is much appreciated !
You can't actually do something like this, because setting 'self' to nil only has any effect within the scope of that method (in your case, 'nilify'). You don't have any actual way to effect the values of pointers located on other parts of the stack or in random places in the heap, for example.
Basically any code that holds a reference to some object is responsible for maintaining and clearing those references itself. If you have some use case where random sections of code may need references to "live" objects of some kind, but where you'd want those object references to go away in response to some external event (maybe a user tracking system or something), you could do something with notifications, but the various modules tracking those "live" objects would still be responsible for listening for notifications and cleaning up references when they received them.
The 'nilify' thing, however, can't possibly work.
You cannot do what you're trying to do. self is just a local reference to an object that actually exists elsewhere. Setting it to nil doesn't mean anything. An object doesn't, in general, own itself, and it certainly doesn't control other objects' references to it. It's up to the owning objects to manage its lifetime.
There are a few things wrong with your code.
First, by convention, class names start with an uppercase letter. Please stick to these naming conventions as it will make it harder for other developers to work with your code (and even confuse you).
Next, your initWithName:... According to the naming conventions, a method with init in its name should be an instance method, not a class method. So either name it newWithName: or turn it into an instance method like this:
-(testA*)initWithA:(NSString *)aString{
self = [super init];
if (!self) return nil;
tst.a=aString;
return self;
}
If you keep it as class method (and name it newWithName:) you should not return a autoreleased object since according to the naming conventions method that start with init... or new... return a retained object. If you do not follow these conventions, the static analyzer will give you "false" warnings and it will become useless for you.
Now for the reason your nillify doesn't work: the self is in fact an argument to a method. Under the hood, your nillify method actually has two arguments that you do not see: the self pointer and the selector pointer. This means, self is actually a variable on the stack. And if you overwrite it, you only overwrite that stack variable but that doesn't influence your test variable which is somewhere else.
As an example, consider a method - (void)foo:(NSString *)bar;. The compiler turns it into the equivalent of the C function (void) foo(id self, SEL _cmd, NSString *bar).
I know this is quite possibly a lame question, but I've pulled three consecutive all-nighters and I'm very blurry. And I'm new to Objective C and Cocoa Touch.
I've created a class that provides a delegate method. I'll use simplified example code since the specifics aren't important. The header file looks like this:
#import <Foundation/Foundation.h>
#protocol UsernameCheckerDelegate <NSObject>
#required
- (void)didTheRequestedThing:(BOOL)wasSuccessful;
#end
#interface TheDelegateClass : NSObject {
id <TheDelegateClassDelegate> tdcDelegate;
}
#property (assign) id <TheDelegateClassDelegate> tdcDelegate;
- (void)methodThatDoesSomething:(int)theValue;
#end
And the source file looks like this:
#import "TheDelegateClass.h"
#implementation TheDelegateClass
#synthesize tdcDelegate;
- (void)methodThatDoesSomething:(int)theValue {
if (theValue > 10) {
[[self tdcDelegate] didTheRequestedThing:NO];
// POINT A
}
// POINT B
int newValue = theValue * 10;
NSString *subject = [NSString stringWithFormat:#"Hey Bob, %i", newValue];
// Some more stuff here, send an email or something, whatever
[[self tdcDelegate] didTheRequestedThing:YES];
// POINT C
}
#end
Here's my question: if theValue is in fact greater than 10 and the line above POINT A runs, does program flow control pass out of this method (and back to the didTheRequestedThing delegate method in the object that called this) or does flow continue on through POINT B to POINT C?
I'm hoping for the former because I can simplify the heck out of my code, currently an unpleasant mess of deeply nested ifs and elses.
When the -didTheRequestedThing: method returns, control flow returns back to your POINT A and continues on to POINT B and POINT C. Delegate method calls are exactly like any other method call. If you want to avoid executing the rest of the method after the delegate call, just stick a call to return where your // POINT A comment is.
I'm doing the following:
- (void) accelerometer: (UIAccelerometer *)accelerometer didAccelerate: (UIAcceleration *)acceleration {
if (self.lastAcceleration) {
double i = self.lastAcceleration.x;
It works fine until I actually tilt the phone. Then I get EXC_BAD_ACCESS on the last line. lastAcceleration is a property with a retain. When I look at "x" in the debugger, it has a large negative value. Why would that throw a EXC_BAD_ACCESS exception only on tilt?
-- EDIT (Since this answer applies to responses below) --
I added this and now it works:
- (void)dealloc {
[lastAcceleration release];
Why would that matter? Also, should it be
[self.lastAcceleration release];
I wasn't previously releasing lastAcceleration anywhere. Here is the header declaration:
#interface MyViewController : UIViewController <UIAccelerometerDelegate> {
UIAcceleration *lastAcceleration;
}
#property(nonatomic, retain) UIAcceleration *lastAcceleration;
#end
My hunch is that the accelerometer API has nothing to do with the crash, the code you have shown smells like bad memory management, given that you're mixing ivar and property access I suspect you might be doing the same in other parts you're not showing.
Anyway a couple best practice things:
any object you have a pointer for in your class you should have retained, and conversely when you release it you should also zap the pointer so you don't risk accessing it after it has been deallocated (the exception to this rule are some patterns like the delegate object, where retaining the object would cause a retain cycle, but that's a whole other topic)
ivar setters and getters that are automatically generated via the #synthesized directive will retain and release the object for you for code that simply looks like it's assigning a pointer, so they're pretty handy, but property access (self.something = ...) and ivar access (something = ...) are not equivalent so you have to be careful
One easy way to make sure you don't mix the two up is to do something like this:
#interface MyObject : NSObject
{
SomethingObject *_something;
}
#property (nonatomic, retain) SomethingObject *something;
#end
#implementation MyObject
#synthesize something = _something;
#end
What we're doing here is making the ivar and property names slightly different, so that you are more aware of which one you're using, and the compiler will bark if you use don't use the bare something = ... syntax.
Now the #synthesize'd accessors are something like this:
- (void)setSomething:(SomethingObject *)newSomething
{
[newSomething retain];
[_something release];
_something = newSomething;
}
- (SomethingObject *)something
{
return _something;
}
With all that out of the way, [lastAcceleration release] is a bad thing to do because it isn't also setting the lastAcceleration pointer to nil, you are not guaranteed that it won't be deallocated and if you accidentally use it you are likely to crash.
[self.lastAcceleration release]; is incorrect because accessors take care of all the retain/release stuff for you.
The correct thing to do here is self.lastAcceleration = nil; that, if you look at the accessor code, will release and set the pointer to nil.
What is likely happening is that you are releasing lastAcceleration somewhere without also setting it to nil, and the if (self.lastAcceleration) { check is hitting a released object.
Main reason to have retained properties is to avoid explicit retain/release calls and memory management bugs associated with them. But in dealloc method either way is fine, since object will cease to exist soon.
[self.lastAcceleration release]; - not necessary.
[lastAcceleration release]; self.lastAcceleration = nil;
Both are fine if used in dealloc.
Outside of dealloc use only
self.lastAcceleration = nil;
EXC_BAD_ACCESS is raised when you access released memory. My guess would be that you somewhere released self.lastAcceleration but didn't set it to null.
Are you sure it is related to tilting?