Question

@implementation ProductController


NSString *areaName = nil;
+ (void)setAreaName:(NSString *)areaName_ {
    areaName = areaName_;
}
@end

and

@implementation ProductController


NSString *areaName = nil;
+ (void)setAreaName:(NSString *)areaName_ {
    if(areaName_ != areaName) {
        [areaName release];
        areaName = [areaName_ copy];
    }
}
- (void)dealloc {
     [areaName release];
}
@end

Now which one is correct?and why?

Answer 1

As you seem to understand, there are no “class variables” in Obj-C. The workaround is just a C-style (global, or file-scoped) variable that you set up similarly to how you’ve shown above. First off, you should use file scope for these variables by marking them with the static keyword:

static NSString *areaName = nil;

You might also consider using a convention like FirstLetterUppercase to indicate the scope difference.

As for memory management, you can treat it exactly like an instance variable, but one that never goes away forever:

static NSString *AreaName = nil;
+ (void)setAreaName:(NSString *)name {
    if (![name isEqualToString:AreaName]) {
        [AreaName release];
        AreaName = [name copy];
    }
}

Note that in your second example, you should NOT release the “class” variable from an instance’s -dealloc method. If you have more than one instance of the object, this leaves a bad dangling pointer, and defeats the purpose of the “class” variable anyways. Generally, when you use this pattern, you’ll “leak” (for some definition of leak) the class variable value, and that’s OK.