JdkObsolete
Suggests alternatives to obsolete JDK classes.

The problem

Some JDK APIs are obsolete and have preferred alternatives.

LinkedList

LinkedList almost never out-performs ArrayList or ArrayDeque¹. If you are using LinkedList as a list, prefer ArrayList. If you are using LinkedList as a stack or queue/deque, prefer ArrayDeque.

Migration gotcha: LinkedList permits null elements; ArrayDeque rejects them. The documentation for Deque strongly discourages users from inserting null, even into implementations that permit it. So, if you are using a LinkedList for this purpose, you should likely stop, and you will need to stop in order to migrate to ArrayDeque.

Vector

Vector performs synchronization that is usually unnecessary; prefer ArrayList.

If a synchronized collection is necessary, use Collections.synchronizedList or a data structure from java.util.concurrent.

Hashtable and Dictionary

This is a nonstandard class that predates the Java Collections Framework; prefer LinkedHashMap or HashMap.

If synchronization is necessary, java.util.concurrent.ConcurrentHashMap is usually a good choice.

java.util.Stack

Stack is a nonstandard class that predates the Java Collections Framework; prefer ArrayDeque.

If a synchronized collection is necessary, use a data structure from java.util.concurrent.

When migrating from Stack to Deque, note that the Stack methods push/pop/peek/add/iterator correspond to the Deque methods addFirst/removeFirst/peekFirst/addFirst/descendingIterator.

StringBuffer

StringBuffer performs synchronization that is rarely necessary and has significant performance overhead. Prefer StringBuilder, which does not do synchronization.

If synchronization is necessary, consider creating an explicit lock object and using synchronized blocks.

Enumeration

An ancient precursor to Iterator.

SortedSet and SortedMap

Replaced by NavigableSet and NavigableMap in Java 6.

Matcher#hitEnd() and Matcher#requireEnd()

The methods [Matcher#hitEnd()][hitEnd] and [Matcher#requireEnd()][requireEnd] are mainly intended for implementing scanning functionalities (like java.util.Scanner), where you need to decide whether to read more input from a stream to find a longer match.

If you are not implementing your own scanning or streaming parser, and the input being matched against is already fully loaded in memory, using these methods is likely a mistake.

Check End of Match

If you want to check if a match extends to the end of the input, you can simply compare the match end index with the length of the input string:

// Instead of:
if (matcher.find() && matcher.hitEnd()) { ... }

// Use:
if (matcher.find() && matcher.end() == input.length()) { ... }

Use java.util.Scanner

If you are implementing a scanner or streaming parser, consider using java.util.Scanner directly, which already encapsulates this logic correctly.

Scanner scanner = new Scanner(input);
while (scanner.hasNext(pattern)) {
  String match = scanner.next(pattern);
  ...
}

Suppression

Suppress false positives by adding the suppression annotation @SuppressWarnings("JdkObsolete") to the enclosing element.