.NET loop integrity 101

Question

I've always been confused about this one. Consider the following loops:

int [] list = new int [] { 1, 2, 3 };  
for (int i=0; i < list.Length; i++) { }  
foreach (int i in list) { }  
while (list.GetEnumerator().MoveNext()) { } // Yes, yes you wouldn't call GetEnumerator with the while. Actually never tried that.  

• The [list] above is hard-coded. If the list was changed externally while the loop was going through iterations, what would happen?
• What if the [list] was a readonly property e.g. int List{get{return(new int [] {1,2,3});}}? Would this upset the loop. If not, would it create a new instance in each iteration?

Answer 1

Well:

• The for loop checks against list.Length on each iteration; you don't actually access list within the loop, so the contents would be irrevelant
• The foreach loop only uses list to get the iterator; changing it to refer to a different list would make no difference, but if you modified the list itself structurally (e.g. by adding a value, if this were really a List<int> instead of an int[]), that would invalidate the iterator
• Your third example as written would go on forever unless the list were cleared, given that it'll get a new iterator each time. If you want a more sensible explanation, post more sensible code

Fundamentally you need to understand the difference between the contents of an array vs changing which object a variable refers to - and then give us a very concrete situation to explain. In general though, a foreach loop only directly touches the source expression once, when it fetches the iterator - whereas a for loop has no magic in it, and how often it's accessed simply depends on the code - both the condition and the "step" part of the for loop are executed on each iteration, so if you refer to the variable in either of those parts, you'll see any changes...