.Net developer new to Objective-C. Need some critique and suggestions - objective-c

I've created one of my first apps using Objective-C. Being a noob, there's a lot of stuff I want to do, but don't know how to apply it in Objective-C. Please take a look at the method below (which I created from scratch) and tell me what you would do to make it better. Obviously I've duplicated code across 2 UILabels, but I'd like to simplify that (I hate duplicating code) but I'm unaware what the best way to do it is. I just need suggestions which will help me better understand the right way to do stuff in Objective-C
timeText and dateText are of type UILabel
- (void)touchesMoved:(NSSet *)touches withEvent:(UIEvent *)event
{
if (isRearranging)
{
NSLog(#"touchesMoved");
NSLog(#"touches=%#,event=%#",touches,event);
//TOUCH INFO
UITouch *touch = [[touches allObjects] objectAtIndex:0];
CGPoint currentLocation = [touch locationInView:touch.view];
CGPoint previousLocation = [touch previousLocationInView:touch.view];
//FRAME INFO
float timeHalfWidth = timeText.frame.size.width / 2;
float timeHalfHeight = timeText.frame.size.height / 2;
CGRect timeTextRect = CGRectMake(timeText.center.x - (timeHalfWidth), timeText.cener.y - (timeHalfHeight), timeText.frame.size.width, timeText.frame.size.height);
float dateHalfWidth = dateText.frame.size.width / 2;
float dateHalfHeight = dateText.frame.size.height / 2;
CGRect dateTextRect = CGRectMake(dateText.center.x - (dateHalfWidth), dateText.center.y - (dateHalfHeight), dateText.frame.size.width, dateText.frame.size.height);
//IF TIME TEXT
if(CGRectContainsPoint(timeTextRect,previousLocation))
{
CGPoint item = timeText.center;
CGPoint diff;
diff.x = previousLocation.x - item.x;
diff.y = previousLocation.y - item.y;
CGPoint newLoc;
newLoc.x = currentLocation.x - diff.x;
newLoc.y = currentLocation.y - diff.y;
if (newLoc.x<timeHalfWidth)
newLoc.x = timeHalfWidth;
if (newLoc.y<timeHalfHeight)
newLoc.y = timeHalfHeight;
[timeText setCenter:(newLoc)];
}
//IF DATE TEXT
if(CGRectContainsPoint(dateTextRect,previousLocation))
{
CGPoint item = dateText.center;
CGPoint diff;
diff.x = previousLocation.x - item.x;
diff.y = previousLocation.y - item.y;
CGPoint newLoc;
newLoc.x = currentLocation.x - diff.x;
newLoc.y = currentLocation.y - diff.y;
if (newLoc.x<dateHalfWidth)
newLoc.x = dateHalfWidth;
if (newLoc.y<dateHalfHeight)
newLoc.y = dateHalfHeight;
[dateText setCenter:(newLoc)];
}
}
touchMoved = YES;
}
Thanks so much for your help!

A first step, independent from the language you are working in, would be to follow DRY - most of your code is the same for both labels.
Then there is already functionality for hit-testing in the SDK, e.g. -hitTest:withEvent: or -pointInside:withEvent::
NSArray *labels = [NSArray arrayWithObjects:timeText, dateText, nil];
for (UILabel *label in labels) {
if ([label pointInside:previousLocation withEvent:nil]) {
[self relocateLabel:label];
break;
}
}

I will answer an additional question posed by asker in a comment. Quote:
What if I wanted to use mixed types in that array? i.e. a couple UILabel and a couple UIImageView. Is there a way to compare types or use generics? This is a poor guess/example for (NSObject *obj in objects) { if (label.type==UILabel) [self relocateLabel:obj]; else if (label.type==UIImageView) [self relocateImage:obj]; }
As Georg Fritzsche answered there, Objective-C messaging is dynamic. Object will be "queried" if it supports that message at runtime, and if so, it will execute the method associated with the message. Method/message name is called a "selector".
If you explicitly want to figure out object's class, you can do that as well.
if([view isKindOfClass:[UILabel class]])
{
// your code here
}
If you just want to figure out if the target object responds to a selector (that is, implements a method):
if([view respondsToSelector:#selector(relocateView:)])
{
// your code here
}
Selectors are derived from method names by omitting arguments themselves, leaving colons intact and appending everything closely. For example, if you had a message send (that is, method call): [thing moveTowardsObject:door movementType:XYZ_CRAWL], its selector would be: moveTowardsObject:movementType: and you'd get it using #selector(moveTowardsObject:movementType:).
In a loop such as what Georg posted, you typically want to just check if the target object responds to a selector, since otherwise an exception would be thrown, and Objective-C code rarely catches exceptions as part of a normal code flow (as opposed to what Python developers do).

Related

Updating NSView asynchronously from a thread

First of all, I'm an Objective-C novice. So I'm not very familiar with OS X or iOS development. My experience is mostly in Java.
I'm creating an agent-based modeling-framework. I'd like to display the simulations and to do that I'm writing a little application. First, a little bit about the framework. The framework has a World class, in which there is a start method, which iterates over all agents and has them perform their tasks. At the end of one "step" of the world (i.e., after all the agents have done their thing), the start method calls the intercept method of an object that implements InterceptorProtocol. This object was previously passed in via the constructor. Using the interceptor, anyone can get a hook into the state of the world. This is useful for logging, or in the scenario that I'm trying to accomplish: displaying the information in a graphical manner. The call to intercept is synchronous.
Now as far as the GUI app is concerned, it is pretty simple. I have a controller that initializes a custom view. This custom view also implements InterceptorProtocol so that it can listen in, to what happens in the world. I create a World object and pass in the view as an interceptor. The view maintains a reference to the world through a private property and so once I have initialized the world, I set the view's world property to the world I have just created (I realize that this creates a cycle, but I need a reference to the world in the drawRect method of the view and the only way I can have it is if I maintain a reference to it from the class).
Since the world's start method is synchronous, I don't start the world up immediately. In the drawRect method I check to see if the world is running. If it is not, I start it up in a background thread. If it is, I examine the world and display all the graphics that I need to.
In the intercept method (which gets called from start running on the background thread), I set setNeedsToDisplay to YES. Since the start method of the world is running in a separate thread, I also have a lock object that I use to synchronize so that I'm not working on the World object while it's being mutated (this part is kind of janky and it's probably not working the way I expect it to - there are more than a few rough spots and I'm simply trying to get a little bit working; I plan to clean up later).
My problem is that the view renders some stuff, and then it pretty much locks up. I can see that the NSLog statements are being called and so the code is running, but nothing is getting updated on the view.
Here's some of the pertinent code:
MasterViewController
#import "MasterViewController.h"
#import "World.h"
#import "InfectableBug.h"
#interface MasterViewController ()
#end
#implementation MasterViewController
- (id)initWithNibName:(NSString *)nibNameOrNil bundle:(NSBundle *)nibBundleOrNil
{
self = [super initWithNibName:nibNameOrNil bundle:nibBundleOrNil];
if (self) {
_worldView = [[WorldView alloc] init];
World* world = [[World alloc] initWithName: #"Bhumi"
rows: 100
columns: 100
iterations: 2000
snapshotInterval: 1
interceptor: _worldView];
for(int i = 0; i < 999; i++) {
NSMutableString* name = [NSMutableString stringWithString: #"HealthyBug"];
[name appendString: [[NSNumber numberWithInt: i] stringValue]];
[world addBug: [[InfectableBug alloc] initWithWorld: world
name: name
layer: #"FirstLayer"
infected: NO
infectionRadius: 1
incubationPeriod: 10
infectionStartIteration: 0]];
}
NSLog(#"Added all bugs. Going to add infected");
[world addBug: [[InfectableBug alloc] initWithWorld: world
name: #"InfectedBug"
layer: #"FirstLayer"
infected: YES
infectionRadius: 1
incubationPeriod: 10
infectionStartIteration: 0]];
[_worldView setWorld: world];
//[world start];
}
return self;
}
- (NSView*) view {
return self.worldView;
}
#end
WorldView
#import "WorldView.h"
#import "World.h"
#import "InfectableBug.h"
#implementation WorldView
#synthesize world;
- (id) initWithFrame:(NSRect) frame {
self = [super initWithFrame:frame];
if (self) {
// Initialization code here.
}
return self;
}
- (void) drawRect:(NSRect) dirtyRect {
CGContextRef myContext = [[NSGraphicsContext currentContext] graphicsPort];
CGContextClearRect(myContext, CGRectMake(0, 0, 1024, 768));
NSUInteger rows = [world rows];
NSUInteger columns = [world columns];
NSUInteger cellWidth = 1024 / columns;
NSUInteger cellHeight = 768 / rows;
if([world running]) {
#synchronized (_lock) {
//Ideally we would need layers, but for now let's just get this to display
NSArray* bugs = [world bugs];
NSEnumerator* enumerator = [bugs objectEnumerator];
InfectableBug* bug;
while ((bug = [enumerator nextObject])) {
if([bug infected] == YES) {
CGContextSetRGBFillColor(myContext, 128, 0, 0, 1);
} else {
CGContextSetRGBFillColor(myContext, 0, 0, 128, 1);
}
NSLog(#"Drawing bug %# at %lu, %lu with width %lu and height %lu", [bug name], [bug x] * cellWidth, [bug y] * cellHeight, cellWidth, cellHeight);
CGContextFillRect(myContext, CGRectMake([bug x] * cellWidth, [bug y] * cellHeight, cellWidth, cellHeight));
}
}
} else {
[world performSelectorInBackground: #selector(start) withObject: nil];
}
}
- (BOOL) isFlipped {
return YES;
}
- (void) intercept: (World *) aWorld {
struct timespec time;
time.tv_sec = 0;
time.tv_nsec = 500000000L;
//nanosleep(&time, NULL);
#synchronized (_lock) {
[self setNeedsDisplay: YES];
}
}
#end
start method in World.m:
- (void) start {
running = YES;
while(currentIteration < iterations) {
#autoreleasepool {
[bugs shuffle];
NSEnumerator* bugEnumerator = [bugs objectEnumerator];
Bug* bug;
while((bug = [bugEnumerator nextObject])) {
NSString* originalLayer = [bug layer];
NSUInteger originalX = [bug x];
NSUInteger originalY = [bug y];
//NSLog(#"Bug %# is going to act and location %i:%i is %#", [bug name], [bug x], [bug y], [self isOccupied: [bug layer] x: [bug x] y: [bug y]] ? #"occupied" : #"not occupied");
[bug act];
//NSLog(#"Bug has acted");
if(![originalLayer isEqualToString: [bug layer]] || originalX != [bug x] || originalY != [bug y]) {
//NSLog(#"Bug has moved");
[self moveBugFrom: originalLayer atX: originalX atY: originalY toLayer: [bug layer] atX: [bug x] atY: [bug y]];
//NSLog(#"Updated bug position");
}
}
if(currentIteration % snapshotInterval == 0) {
[interceptor intercept: self];
}
currentIteration++;
}
}
//NSLog(#"Done.");
}
Please let me know if you'd like to see any other code. I realize that the code is not pretty; I was just trying to get stuff to work and I plan on cleaning it up later. Also, if I'm violating an Objective-C best practices, please let me know!
Stepping out for a bit; sorry if I don't respond immediately!
Whew, quiet a question for probably a simple answer: ;)
UI updates have to be performed on the main thread
If I read your code correctly, you call the start method on a background thread. The start method contains stuff like moveBugFrom:... and also the intercept: method. The intercept method thus calls setNeedsDisplay: on a background thread.
Have all UI related stuff perform on the main thread. Your best bet is to use Grand Central Dispatch, unless you need to support iOS < 4 or OS X < 10.6 (or was it 10.7?), like this:
dispatch_async(dispatch_get_main_queue(), ^{
// perform UI updates
});

exc bad access error unless value is hard coded

I am setting the value of a key in an NSDisctionary using this:
- (void)touchesEnded:(NSSet *)touches withEvent:(UIEvent *)event {
NSLog(#"touchesEnded currentcolor: %#", currentColor);
if (currentTool == #"paint") {
NSLog(#"End Paint");
CGPoint point = [[touches anyObject] locationInView:self.view];
MaskPath * myNewPath = [mapView.drawthis containsPoint:point];
[myNewPath.attributes setValue:currentColor forKey:#"fill"];
[mapView setNeedsDisplay];
} else {
NSLog(#"End Pattern");
currentColor = #"1.png";
CGPoint point = [[touches anyObject] locationInView:self.view];
MaskPath * myNewPath = [mapView.drawthis containsPoint:point];
[myNewPath.attributes setValue:currentColor forKey:#"fill"];
[mapView setNeedsDisplay];
}
}
If I try to log the value of currentColor the app crashes with exc bad access (line 3)
If I take the Log out and go with the hardcoded value everything works fine. It also works fine with the first part of the if statement. I've checked the function that assigns currentColor and it is delivering the correct values. If I hard code the value of currentColor at that point it works fine. I've run the analyzer and I've got no memory leaks or issues. How else can I track this issue down?
One glaring issue is that you are working with an autorelease variable as evidenced by your example. Most likely you are doing the same thing outside of that method.
I recommend you make currentColor a property:
#property (nonatomic, retain) NSString *currentColor;
Then, to set it, you would do something like this:
self.currentColor = #"SomeColor";
Since you declared the currentColor property with a retain, you will not get a bad access when you pass it to your NSLog:
NSLog(#"This is my color: %#", self.currentColor);
If you already DO have a currentColor property and you are simply assigning the ivar, then you should be doing:
currentColor = [NSString alloc] initWithString:#"Foo"];
To sum it up:
1. self.currentColor = #"Foo"; //<-This is automagically retained
2. currentColor = [NSString alloc] initWithString:#"Foo"]; //<-Accessing the ivar directly, so you must ownify the string :D

Handling Double Tap in Cocos2d

This question is too primitive, but I have been trying for the last 8 hours, and it is eating away my energy ( and confidence levels too :))
In my class, I have registered for Targeted touches.
[[CCTouchDispatcher sharedDispatcher] addTargetedDelegate:self priority:-1 swallowsTouches:YES];
I would like to check if the user is performing single tap or double tap. In ccTouchBegan method, first gets a touch with singleTap, then gets a touch with doubleTap. I found a very interesting solution for this one, in one of the Cocos2d Forums. ( I could not locate it now.. )
The solution goes like this.
switch (touch.tapCount) {
case 2:
[NSObject cancelPreviousPerformRequestsWithTarget:self selector:#selector(handleSingleTouch:) object:singleTouchLocation];
// Handle Double Touch here
break;
case 1:
self.singleTapLocation = ccp(location.x,location.y);
self.singleTouchLocation = touch;
[self performSelector:#selector(handleSingleTouch:) withObject:touch afterDelay:0.3
];
break;
}
Everything looks to be working fine. Essentially, you schedule a single touch handler with a delay and see if it is actually a double tap. If it is a double tap, then cancel singletap handler and perform logic for double tap.
But my problem is, In the single touch handler (handleSingleTouch), I am adding a CCSprite to the UI. This action is not working. No sprite gets added. (Although the method is called.). In fact this works, If I call the selector with no delay, but with delay, the sprite is not added.
I am not a Objective C expert, So, I apologize if the question is too primitive.
Edit 1: The original thread.
Edit 2: Posting handleSingleTouch.. only relevant code.
-(void) handleSingleTouch:(UITouch * ) touch {
CGPoint location = [touch locationInView: [touch view]];
CGPoint glLocation = [MainSceneLayer locationFromTouch:touch];
//Single tap
CCLOG(#"Single tap: Adding marker image");
if(zoomedOut) {
CGPoint globalCoordinates = [quadConvertor convertLocalToGlobal:location
inActiveQuadrant:activeQuadrant];
if( [self isTouchValidForSelectedItem:globalCoordinates] == Matched) {
[self addMarkerImages:glLocation];
} else if([self isTouchValidForSelectedItem:globalCoordinates] == NotMatched) {
[missedLabel stopAllActions];
for (ImageQuadrants* quad in quadrants) {
if(quad.quadrantNumber == activeQuadrant) {
missedLabel.position = ccp((quad.center.x * -1)+glLocation.x , (quad.center.y * -1)+glLocation.y);
//markers.position = ccp((quad.center.x * -1)+touchPosition.x , 320);
}
}
id blinkAction = [CCBlink actionWithDuration:1 blinks:3];
id makeInvible = [CCCallFunc actionWithTarget:self selector:#selector(makeInvisible:)];
id seq = [CCSequence actions:blinkAction, makeInvible, nil];
[missedLabel runAction:seq];
} else {
[alreadyAccountedLabel stopAllActions];
for (ImageQuadrants* quad in quadrants) {
if(quad.quadrantNumber == activeQuadrant) {
alreadyAccountedLabel.position = ccp((quad.center.x * -1)+glLocation.x , (quad.center.y * -1)+glLocation.y);
//markers.position = ccp((quad.center.x * -1)+touchPosition.x , 320);
}
}
id blinkAction = [CCBlink actionWithDuration:1 blinks:3];
id makeInvible = [CCCallFunc actionWithTarget:self selector:#selector(makeInvisible1:)];
id seq = [CCSequence actions:blinkAction, makeInvible, nil];
[alreadyAccountedLabel runAction:seq];
}
}
swipeStartPoint = [touch locationInView:touch.view];
}
i dont know if thats a typo in your question only but your delay method is all wrong. There is no withDelay: argument. it should look like this:
[self performSelector:#selector(handleSingleTouch:) withObject:touch afterDelay:0.1];
EDIT:
Since your problem is most probably with the touch getting lost. Try [touch retain] before calling the delayed method. Another way is changing the handleSingleTouch: method to take two floats x and y or a CCPoint as arguments instead of a UITouch. Then you create the floats before the delayed method and pass them in the delayed method. This way you will avoid a memory leak from retaining a touch as well since most probably you cant release it after calling the delayed method.
hope this helps

CGPointMake(random(), random()) not working [duplicate]

This question already has answers here:
Closed 11 years ago.
Possible Duplicate:
Generating Random Numbers in Objective-C
Hi I am creating an app where when you press a pimple it travels to a random place in the view. THe pimple is a UIImageView. Here is the code that I put in:
-(void)touchesBegan:(NSSet *)touches withEvent:(UIEvent *)event
{
UITouch *myTouch = [touches anyObject];
CGPoint point = [myTouch locationInView:pimple];
if ( CGRectContainsPoint(pimple.bounds, point) ) {
[self checkcollision];
}
}
-(void)checkcollision
{
pimple.center = CGPointMake(random(), random());
sad.hidden = NO;
NSURL* popURL = [NSURL fileURLWithPath:[[NSBundle mainBundle] pathForResource:#"pop" ofType:#"mp3"]];
AudioServicesCreateSystemSoundID((CFURLRef) popURL, &popID);
AudioServicesPlaySystemSound(popID);
}
This code: pimple.center = CGPointMake(random(), random()); does not work. When I press the pimple, the pimple dissapears. Since I have not stated any .hidden for the pimple it must mean it IS moving. But you cant see it, (its outside the view.) Is there something wrong with my code? Am i missing something? Please help.
afaik random() ist just the same as rand() in c which
Returns a pseudo-random integral number in the range 0 to RAND_MAX. [...] RAND_MAX is a constant defined in cstdlib. Its default value may vary between implementations but it is granted to be at least 32767.
As a result you move our pimple to a random position inside a rect that is much bigger than your window. Try this: (I assume that self is a UIView)
pimple.center = CGPointMake(
random() % (unsigned int)self.bounds.size.width,
random() % (unsigned int)self.bounds.size.height);
The %-Operator is the Modulo-Operator.
It seems that you dont seed your pseudo-random-generator which leeds to always the same trajectory on each startup. There is another Thread on stackoverflow which tells us to use arc4random instead of random()/rand()

Eliminate Pin Overlap in the MKMapView

I am working with MKMapView Based application. I need a clarification whether it is possible to eliminate the pin OverLap in the MKMapView? Because at some places there are large number of pins displaying. It is difficult to me to identify the pins.
If you have an Apple Developer Account, I would strongly recommend getting the Session 111 video from the 2011 WWDC Conference Sessions, entitled "Visualizing Information Geographically with MapKit". One of the segments specifically covers how to cluster content from large data sets to allow you to group or ungroup pins based on density at various zoom levels.
Their example is elegantly simple, but at the heart of the problem you want to replace a group of overlapping pins with a single pin and as you zoom in the single pin will split back into the individual pins.
How and when you decide to group things can be varied considerably. Apple's solution simply subdivides the map into a grid and any box that has more than 1 pin results in a group. You could also take an algorithmic approach such as using a kMeansCluster algorithm which is incredibly simple and you could feed all of your annotations through the algorithm and get an array of groups out the other side logically organized.
From there it's a matter of keeping track of all the individual pins and how they are grouped as you zoom in and out. You will only display a single annotation for each group or any individual pins that are left over. It's also possible to animate the transitions as the map zooms in and out so you can visually reinforce what is happening.
My own technique is too closely related to Apple's approach for me to post here so I'm hoping you can access the above video which covers almost all of these points.
For this you have to implement clustering concept to your map.By using Apple demo code it's easy to implement clustering concept in our code. Reference link
Simply we can use following code for the Clustering
Steps to implement clustering
Step1 : The important thing is for clustering we use two mapviews(allAnnotationsMapView, ), One is for reference(allAnnotationsMapView).
#property (nonatomic, strong) MKMapView *allAnnotationsMapView;
#property (nonatomic, strong) IBOutlet MKMapView *mapView;
In viewDidLoad
_allAnnotationsMapView = [[MKMapView alloc] initWithFrame:CGRectZero];
Step2 : Add all annotations to the _allAnnotationsMapView, In below _photos are the annotations array.
[_allAnnotationsMapView addAnnotations:_photos];
[self updateVisibleAnnotations];
Step3 : Add below methods for clustering, in this PhotoAnnotation is the custom annotation.
MapViewDelegate methods
- (void)mapView:(MKMapView *)aMapView regionDidChangeAnimated:(BOOL)animated {
[self updateVisibleAnnotations];
}
- (void)mapView:(MKMapView *)aMapView didAddAnnotationViews:(NSArray *)views {
for (MKAnnotationView *annotationView in views) {
if (![annotationView.annotation isKindOfClass:[PhotoAnnotation class]]) {
continue;
}
PhotoAnnotation *annotation = (PhotoAnnotation *)annotationView.annotation;
if (annotation.clusterAnnotation != nil) {
// animate the annotation from it's old container's coordinate, to its actual coordinate
CLLocationCoordinate2D actualCoordinate = annotation.coordinate;
CLLocationCoordinate2D containerCoordinate = annotation.clusterAnnotation.coordinate;
// since it's displayed on the map, it is no longer contained by another annotation,
// (We couldn't reset this in -updateVisibleAnnotations because we needed the reference to it here
// to get the containerCoordinate)
annotation.clusterAnnotation = nil;
annotation.coordinate = containerCoordinate;
[UIView animateWithDuration:0.3 animations:^{
annotation.coordinate = actualCoordinate;
}];
}
}
}
clustering Handling methods
- (id<MKAnnotation>)annotationInGrid:(MKMapRect)gridMapRect usingAnnotations:(NSSet *)annotations {
// first, see if one of the annotations we were already showing is in this mapRect
NSSet *visibleAnnotationsInBucket = [self.mapView annotationsInMapRect:gridMapRect];
NSSet *annotationsForGridSet = [annotations objectsPassingTest:^BOOL(id obj, BOOL *stop) {
BOOL returnValue = ([visibleAnnotationsInBucket containsObject:obj]);
if (returnValue)
{
*stop = YES;
}
return returnValue;
}];
if (annotationsForGridSet.count != 0) {
return [annotationsForGridSet anyObject];
}
// otherwise, sort the annotations based on their distance from the center of the grid square,
// then choose the one closest to the center to show
MKMapPoint centerMapPoint = MKMapPointMake(MKMapRectGetMidX(gridMapRect), MKMapRectGetMidY(gridMapRect));
NSArray *sortedAnnotations = [[annotations allObjects] sortedArrayUsingComparator:^(id obj1, id obj2) {
MKMapPoint mapPoint1 = MKMapPointForCoordinate(((id<MKAnnotation>)obj1).coordinate);
MKMapPoint mapPoint2 = MKMapPointForCoordinate(((id<MKAnnotation>)obj2).coordinate);
CLLocationDistance distance1 = MKMetersBetweenMapPoints(mapPoint1, centerMapPoint);
CLLocationDistance distance2 = MKMetersBetweenMapPoints(mapPoint2, centerMapPoint);
if (distance1 < distance2) {
return NSOrderedAscending;
} else if (distance1 > distance2) {
return NSOrderedDescending;
}
return NSOrderedSame;
}];
PhotoAnnotation *photoAnn = sortedAnnotations[0];
NSLog(#"lat long %f %f", photoAnn.coordinate.latitude, photoAnn.coordinate.longitude);
return sortedAnnotations[0];
}
- (void)updateVisibleAnnotations {
// This value to controls the number of off screen annotations are displayed.
// A bigger number means more annotations, less chance of seeing annotation views pop in but decreased performance.
// A smaller number means fewer annotations, more chance of seeing annotation views pop in but better performance.
static float marginFactor = 2.0;
// Adjust this roughly based on the dimensions of your annotations views.
// Bigger numbers more aggressively coalesce annotations (fewer annotations displayed but better performance).
// Numbers too small result in overlapping annotations views and too many annotations on screen.
static float bucketSize = 60.0;
// find all the annotations in the visible area + a wide margin to avoid popping annotation views in and out while panning the map.
MKMapRect visibleMapRect = [self.mapView visibleMapRect];
MKMapRect adjustedVisibleMapRect = MKMapRectInset(visibleMapRect, -marginFactor * visibleMapRect.size.width, -marginFactor * visibleMapRect.size.height);
// determine how wide each bucket will be, as a MKMapRect square
CLLocationCoordinate2D leftCoordinate = [self.mapView convertPoint:CGPointZero toCoordinateFromView:self.view];
CLLocationCoordinate2D rightCoordinate = [self.mapView convertPoint:CGPointMake(bucketSize, 0) toCoordinateFromView:self.view];
double gridSize = MKMapPointForCoordinate(rightCoordinate).x - MKMapPointForCoordinate(leftCoordinate).x;
MKMapRect gridMapRect = MKMapRectMake(0, 0, gridSize, gridSize);
// condense annotations, with a padding of two squares, around the visibleMapRect
double startX = floor(MKMapRectGetMinX(adjustedVisibleMapRect) / gridSize) * gridSize;
double startY = floor(MKMapRectGetMinY(adjustedVisibleMapRect) / gridSize) * gridSize;
double endX = floor(MKMapRectGetMaxX(adjustedVisibleMapRect) / gridSize) * gridSize;
double endY = floor(MKMapRectGetMaxY(adjustedVisibleMapRect) / gridSize) * gridSize;
// for each square in our grid, pick one annotation to show
gridMapRect.origin.y = startY;
while (MKMapRectGetMinY(gridMapRect) <= endY) {
gridMapRect.origin.x = startX;
while (MKMapRectGetMinX(gridMapRect) <= endX) {
NSSet *allAnnotationsInBucket = [self.allAnnotationsMapView annotationsInMapRect:gridMapRect];
NSSet *visibleAnnotationsInBucket = [self.mapView annotationsInMapRect:gridMapRect];
// we only care about PhotoAnnotations
NSMutableSet *filteredAnnotationsInBucket = [[allAnnotationsInBucket objectsPassingTest:^BOOL(id obj, BOOL *stop) {
return ([obj isKindOfClass:[PhotoAnnotation class]]);
}] mutableCopy];
if (filteredAnnotationsInBucket.count > 0) {
PhotoAnnotation *annotationForGrid = (PhotoAnnotation *)[self annotationInGrid:gridMapRect usingAnnotations:filteredAnnotationsInBucket];
[filteredAnnotationsInBucket removeObject:annotationForGrid];
// give the annotationForGrid a reference to all the annotations it will represent
annotationForGrid.containedAnnotations = [filteredAnnotationsInBucket allObjects];
[self.mapView addAnnotation:annotationForGrid];
for (PhotoAnnotation *annotation in filteredAnnotationsInBucket) {
// give all the other annotations a reference to the one which is representing them
annotation.clusterAnnotation = annotationForGrid;
annotation.containedAnnotations = nil;
// remove annotations which we've decided to cluster
if ([visibleAnnotationsInBucket containsObject:annotation]) {
CLLocationCoordinate2D actualCoordinate = annotation.coordinate;
[UIView animateWithDuration:0.3 animations:^{
annotation.coordinate = annotation.clusterAnnotation.coordinate;
} completion:^(BOOL finished) {
annotation.coordinate = actualCoordinate;
[self.mapView removeAnnotation:annotation];
}];
}
}
}
gridMapRect.origin.x += gridSize;
}
gridMapRect.origin.y += gridSize;
}
}
By following above steps we can achieve clustering on mapview, it is not necessary to use any third party code or framework. Please check the Apple sample code here. Please let me know if you have any doubts on this.
It's quite easy to implement your own annotation clustering framework. Here's an example of a basic one that you can refer here.
If your pins are overlapping then it must be that your zoom level is high for that place.
You can think of removing some annotations in that zoom level until you dont have annotation overlaps and while zooming in you can add the annotations so that there are enough space between the annotations.