I have a class with the following init method:
- (id)init
{
self = [super init];
if (self) {
// Initialization code here.
StateStack* s = [[StateStack alloc] init];
state = s;
[s push:NONE]; //<--EXC_BAD_ACCESS on load here
[s release];
}
return self;
}
And StateStack has the following init code:
- (id)init {
self = [super init];
if (self) {
NSMutableArray* s = [[NSMutableArray alloc] init];
stack = s;
[s release];
NSLog(@"%d",[stack retainCount]);
}
return self;
}
Oddly, if I remove the NSLog line, the EXC_BAD_ACCESS moves to StateStack’s dealloc method:
- (void)dealloc {
[stack release]; //<--EXC_BAD_ACCESS
[super dealloc];
}
Searching around seems to suggest that EXC_BAD_ACCESS is caused by overreleasing, but I can’t see how I’ve overreleased anything. Does anyone know what the cause might be?
In your
initfunction:you are allocating an instance of
StateStack; this gets a retain count of 1. Then at the end of the function you callrelease, retain count goes to 0 and the object is ready to be released. So, when laterdeallocis executed, thestateivar is sent anotherreleaseand that is causing the bad access. You don’t need to release s, since you want that state be retained. The same error pattern occurs in the other init method.This would be correct:
NB: I don’t want to generate misunderstandings. Using retain count to check for correct memory allocation is useless. This is true. Anyway, reasoning in terms of retain count helps understanding what happens when you allocate/release/autorelease an object. It is the basic mechanism, but it is too difficult to track it usage to check for correctness of memory management.