Retaining repeating NSTimer for later access?

Question

I am creating an NSTimer in the createTimer method that I want to refer back to in the later cancelTimer method. To facilitate this I am taking ownersh

Answer 1

You need to invalidate before you release. After the timer has fire, you are the only one holding a retain on the timer. So when you call release, the timer deallocates. You then call invalidate on invalid memory and you crash.

Answer 2

You release before you call invalidate. That means by the time you call invalidate, you've already relinquished ownership of the timer. In practice, you end up calling invalidate on a deallocated timer instance.

What you should do is call invalidate before you call release. Since you are using a retained property, you can just set the property to nil:

// Schedule the timer.
self.walkTimer = [NSTimer scheduledTimerWith...];

// Cancel the timer.
[self.walkTimer invalidate];
self.walkTimer = nil;

---

Update to clear up any confusion regarding memory management

It's important to keep in mind the Memory Management Rules of Objective-C — you own an object if you call alloc, copy or retain on it, and if you own an object, you have to eventually call release. In this case, setWalkTimer: retains the timer because the property is declared as retain — that means you own the timer and must call release on it down the road. The invalidate method does not count as relinquishing ownership of the timer.

When you schedule a timer, the run loop retains it, and when the timer fires or is invalidated, the run loop releases it. But really, you don't need to know that — it's an implementation detail. The call to release by invalidate is only to balance the retain when the timer was scheduled on the run loop.

Answer 3

Don't retain a scheduled NSTimer if you've set its target to self Don't set self as the target of a repeating timer unless you are absolutely sure to know all the consequences

(...otherwise the runtime drowns a kitten in leaked timers, targets and userInfos — or so goes the saying.)

Please read and re-read "Overview" in the NSTimer Class Reference and pay special attention to the last paragraph.

In a nutshell:

1. If you schedule an NSTimer, it becomes associated to the current run-loop which retains it.
2. Furthermore, the timer retains its target.
3. NSTimer instances are not reusable: "Once invalidated, timer objects cannot be reused".

So there is no point in retaining a scheduled timer in the first place.

If you need to hang on to it (e.g. in order to cancel it) use a non-owning (aka weak) reference to it.

Update:
For a thorough explanation, see my answer to your other question (it now has graphs — albeit as links only — and stuff).

Please consider the rest of this post (as well as many of my comments) as obsolete.

---

Your property becomes

@property (nonatomic, assign) NSTimer *walkTimer;

BTW:

-(void)cancelTimer {
    [self setWalkTimer:nil]; // great, now [self walkTimer] returns nil so
    [[self walkTimer] invalidate]; // here, you are calling [nil invalidate]
}

And since messaging nil is absolutely fine in Objective C, your crash miraculously vanishes...while your timer will happily continue to fire.

Edit

I've forgot to mention:
A timer wants a selector that takes one argument, which will be the timer that fired... Or is that just a typo?