Page History

            According to the Java API documentation \[[API 2006|AA. Bibliography#API 06]\] for the {{Iterator.remove()}} method:

The behavior of an iterator is unspecified if the underlying collection is modified while the iteration is in progress in any way other than by calling this method.

...

According to the Java API documentation \[[API 2006|AA. Bibliography#API 06]\] for {{ConcurrentModificationException}}:

... it is not generally permissible for one thread to modify a Collection while another thread is iterating over it. In general, the results of the iteration are undefined under these circumstances. Some Iterator implementations (including those of all the general purpose collection implementations provided by the JRE) may choose to throw this exception if this behavior is detected. Iterators that do this are known as fail-fast iterators, as they fail quickly and cleanly, rather that risking arbitrary, non-deterministic behavior at an undetermined time in the future.

Note that fail-fast behavior cannot be guaranteed because it is, generally speaking, impossible to make any hard guarantees in the presence of unsynchronized concurrent modification. Fail-fast operations throw ConcurrentModificationException on a best-effort basis. Consequently, it would be wrong to write a program that depended on this exception for its correctness: ConcurrentModificationException should be used only to detect bugs.

Reliance on {{ConcurrentModificationException}} is insufficient to stop side effects resulting from modifying an underlying Collection while simultaneously iterating over the collection. The fail-fast behavior may occur only after processing an arbitrary number of elements. In _Java Concurrency In Practice_ \[[Goetz 2006|AA. Bibliography#Goetz 06]\], Göetz et al. note "\[Fail-fast iterators\] are implemented by associating a modification count with the collection: if the modification count changes during iteration, {{hasNext}} or {{next}} throws {{ConcurrentModificationException}}. However, this check is done without synchronization, so there is a risk of seeing a stale value of the modification count and therefore that the iterator does not realize a modification has been made. This was a deliberate design tradeoff to reduce the performance impact of the concurrent modification detection code.".

Note that the enhanced for loop (for-each idiom) uses an Iterator internally. Consequently, enhanced for loops can also participate in concurrent modification issues, even though they lack an obvious iterator.

...

This approach needs to be implemented correctly to avoid starvation, deadlock, and scalability issues \[[Goetz 2006|AA. Bibliography#Goetz 06]\]. 

...

Creating deep copies of the list prevents underlying changes in the original list from affecting the iteration in progress. "Since the clone is thread-confined, no other thread can modify it during iteration, eliminating the possibility of {{ConcurrentModificationException}}. (The collection still must be locked during the clone operation itself)" \[[Goetz 2006|AA. Bibliography#Goetz 06]\]. However, this approach is often more expensive than other techniques. There is also a risk of operating on stale data, which may affect the correctness of the code.

...

The CopyOnWriteArrayList data structure implements all mutating operations by making a fresh copy of the underlying array. It is fully thread-safe , and is optimized for cases where traversal operations vastly outnumber mutations. Note that traversals of such lists always see the list in the state it had at the creation of the iterator (or enhanced for loop); subsequent modifications of the list are invisible to an ongoing traversal. Consequently, this solution is inappropriate when mutations of the list are frequent or when new values should be reflected in ongoing traversals.

...

...