iphone - Assigned value is garbage or undefined logic error in XCode -

I am updating our iPhone app for iOS 4 and I have been involved in a problem "near the function call -By-debit is undefined "in the code

 for  (i = 0; i  

Which is fair enough, the width [ii] has not been started. By updating the code (below), I get this new error, "Assigned value is trash or undefined". Reading on this I think segWidth preserves its value as garbage - is it correct and if so, how can I clean it?

i ; I

------------- Edit ------------------

< P> Thanks for the guys Thanks more info as follows;

A generalized version of the method is shown below as some function. I have removed ugly artists but still looking at logic error for the "Assigned value trash or undefined" line segWidth = width [i];

I believe that this value is seen, there is no clear initiality in the width [i], I am unsure if I lack the understanding of the original obz-c float, or if my assignment There is no logic defect in the syntax?

  - (zero) Some functions {unsigned int n = self.numberOfSegments; Float width [n]; For (i = 0; i & lt; n; i ++) {width [i] = some flight values; } ... for (i = 0; i & lt; self.numberOfSegments; i ++) {float segWidth = 0.0; SegWidth = width [i]; [Self setwidth: Round for Segment (Segvideth): IdeaX: i]; }}  

The definition of setwead is:

  - (zero) setweed: setgate andx for width (CGFlot): (NSUntengar) segment; // Set AutoStage 0.0 Width to Default 0.0  

I agree I'm calling you:

  - (zero) SETview: Segment for width of (CGFLOT) ATX: (NSUntengar) Segment  

And that Pass-by-log is undefined in the function call In fact, LLVM static analyzer error when you build and analyze (rather than significant data points - Are you going through exactly the same way, which were and exactly how useful output).

And you say that width [i] han '

width of segWidth = (float) [i] You should be OK to add a lot more reason to complain with the analyst ** Assigned value is garbage or undefined * *. You have not really set anything to width [i] . I suggest filing a bug against static analyzer, however, because this first error message is actually quite frustrating .

As Joshua also said that cast is really weird, in general, you should use very few types of castings in purpose and never use it on scalar types. should do.

---

Consider two loop calculations:

 {...} for (i = 0; i & lt; n; i ++) {.. (I = 0; i & lt; self.numberOfSegments; i ++) {...}  

Static analyzer does not know that n == self .numberOfSegments And, as such, it should be assumed that the second loop can be more loop than before. Now, you could say, "But I gave up the n = self.numberOfSegments ?!"

You did, but this value could change between the first call and the second call, in this way, the analyst correctly identified that you can use an initial value.

(In fact, it should be said that you can run away from the end of the array, because it is a real risk).