How to use Scanner

The java.util.Scanner class built into java can be used to read inputs. It can for example be used to set up an interactive command line session, where you prompt the user for stuff and your program acts on what the user types in. You can also point it at a file or network connection if you like.

TL;DR: Do not use nextLine()

Phew, glad we got that out there. If you don’t want to read the rest, just never call that method. Instead, to read a full line of text, you call:

scanner.useDelimiter("\r?\n");
String entireLine = scanner.next();

Creating scanners

To create a scanner, just call its constructor, and pass in the data you want to read from. You have lots of options:

import java.util.Scanner;
import java.io.InputStream;
import java.nio.file.Paths;
import java.nio.charset.StandardCharsets;
Scanner in = new Scanner(System.in);
Scanner in = new Scanner(Paths.get("/path/to/file"), StandardCharsets.UTF_8);
Scanner in = new Scanner("input to read");
try (InputStream raw = MyClass.class.getResource("listOfStates.txt")) {
  Scanner in = new Scanner(raw, StandardCharsets.UTF_8);
}

The above code shows 4 different things you can read with scanner:

  1. “Standard input” – what the user types into the command line.
  2. Files. You can also pass a java.io.File object if you still use the old file API.
  3. A string containing data you want to read. The string is read directly; you can’t pass a file path (see the second example for reading files).
  4. Input streams, byte channels, and Readables.

Note that in all cases where you’re dealing with byte input, you can opt not to explicitly pass along a character encoding, but don’t do this! – this means you get the platform default and you have no idea what that’s going to be. That is a good way to write code that works on your computer but fails elsewhere. When in doubt, pass in StandardCharsets.UTF_8 just like in the examples above.

Reading data from a scanner

To read data from a scanner, first decide what you want to read. Then, call the right next method:

A whole number

Call nextInt(), nextLong(), or nextBigInteger(). If you don’t know what these are, call nextInt().

A number in hexadecimal

Call nextInt(0x10), or nextLong(0x10).

A number with fractional elements (i.e., numbers after a decimal point)

Call nextDouble(). However, be aware that how people type floating point numbers depends on where they live. You may want to call useLocale() first to explicitly set the style.

A boolean value

Call nextBoolean()

A single word

Call next()

A whole sentence

Ah, well, that’s tricky. Do not call nextLine(), that doesn’t do what you might think based on its name. Instead, read the next section.

Tokens?

The way Scanner works is by first reading a so-called ‘token’, and then converting this into the representation you asked for. So, if you call nextInt(), a single token is read and this token is then parsed as an integer. If it isn’t an integer, for example because the user typed hello instead of 85, an InputMismatchException is thrown. Catch this exception if you wish to respond to faulty inputs (I advise not calling hasNextInt() and friends either; but that’s for another article).
By default, ‘read the next token’ is done by reading until end-of-input, or any whitespace (tabs, spaces, enters, anything goes). This means that ‘read the next entire line’ just isn’t a thing scanner can do: That would involve reading multiple tokens.
The trick is to turn ‘the entire line’ into a single token. To do that, just tell Scanner to delimit on something else! The delimiter is what separates tokens. So, to read an entire line, first tell scanner that tokens are separated only by newlines, then just read the next token:

scanner.useDelimiter("\r?\n");
String entireLine = scanner.next();

This sets the delimiter to be the newline character, optionally preceded by the carriage return character (windows formatted text files tend to put that in front of their newlines, other OSes don’t).
If you want to return to whitespace-separated tokens, you just call scanner.reset(), though, note, that also resets useLocale.

So what does nextLine do?

nextLine() is unlike all the other next methods. It doesn’t read tokens. Instead, it just keeps reading characters up to the next newline character and returns what it read. It completely ignores the delimiter. This is problematic, because the next method family reads up to the delimiter but doesn’t ‘consume’ it. Thus, if you write this code:

Scanner scanner = new Scanner(System.in);
System.out.print("Enter your age: ");
int age = scanner.nextInt();
System.out.println("Enter your name: ");
String name = scanner.nextLine();

It won’t work: In the user types 5 and then hits enter, well, the 5 goes to nextInt() and that lone enter is all that nextLine is going to read. But if the user types 5 Jane Doe then you’ll end up with age = 5 and name = "Jane Doe". There’s no real way to try to fix this (you can try to call nextLine() twice but that breaks it for those who type spaces instead of enter); just don’t use nextLine, and change the delimiter instead.

How (not) to extend standard collection classes

Today in ##java, someone had a problem: They wanted a bounded-size PriorityQueue. One solution offered was extending the PriorityQueue class to add this behavior1.
This article will talk about why that is not a good solution.
Here is an example implementation of this approach:

public class BoundedPriorityQueue<E> extends PriorityQueue<E> {
    private final int bound;
    public BoundedPriorityQueue(Comparator<? super E> comparator, int bound) {
        super(comparator);
        this.bound = bound;
    }
    @Override
    public boolean add(E e) {
        if (size() >= bound) { return false; }
        return super.add(e);
    }
}

Now… This looks somewhat correct, if you don’t pay attention. But it isn’t! The important thing missed here is that the Queue.offer(E) method is not overridden, and gives the user a way to put this queue into an invalid state (too many items).
In this particular case this is not a huge issue. There are only two methods, offer and add, that increase the size of a PriorityQueue, and add happens to delegate to offer. The important thing to get out of this example is that this bug can happen in the first place, and that the compiler – or runtime – won’t warn you about it.

Extending ArrayList

Let’s get to a more problematic example. We want to have a List that can only be added to, but never removed from.

public class AddOnlyArrayList<E> extends ArrayList<E> {
    @Override public E remove(int index) { throw new UnsupportedOperationException(); }
    @Override public boolean remove(Object o) { throw new UnsupportedOperationException(); }
    @Override public boolean removeAll(Collection<?> c) { throw new UnsupportedOperationException(); }
    @Override public void clear() { throw new UnsupportedOperationException(); }
    @Override public Iterator<E> iterator() { return Iterators.unmodifiableIterator(super.iterator()); }
    @Override public List<E> subList(int fromIndex, int toIndex) { return ...; }
    @Override public ListIterator<E> listIterator() { return ...; }
    @Override public ListIterator<E> listIterator(int index) { return ...; }
}

Hey, this is good, right? All remove methods and clear accounted for, iterators and subList also implemented, should be fine! Well, that’d be right. In Java 7.
With Java 8, Collection.removeIf(Predicate<E>) was added. This is a default method on Collection that normally only delegates to iterator, but ArrayList has its own, optimized version of it.
Suddenly, just with a JDK update, our class breaks and we may not even notice it. The issue is pretty subtle – it may never appear, and when it does happen, the only way we’ll notice is that our “AddOnly”ArrayList suddenly is shrinking because someone elsewhere decided to use that fancy new removeIf method.
The core problem with both these examples is that extension is fragile. A class has to be designed with extension in mind if it’s supposed to be extensible this way. While most standard library classes aren’t final, many of them are not intended to be extended, or only offer hooks for some specific forms of extension.

How to do it properly

You hear this repeated often, but many do not really seem to understand why: Prefer composition over inheritance. Here’s our AddOnlyArrayList with composition:

public class AddOnlyArrayList<E> extends AbstractList<E> {
    private final List<E> elements = new ArrayList<>();
    @Override public E get(int index) { return elements.get(index); }
    @Override public int size() { return elements.size(); }
    @Override public E set(int index, E element) { return elements.set(index, element); }
    @Override public void add(int index, E element) { elements.add(index, element); }
}

remove is unsupported by this class, but importantly, this can never change. The elements list we’re delegating to is fully encapsulated, and we control all access to it. There’s no way for a future change in List, AbstractList or even ArrayList itself to allow removal of elements. This approach is a lot less fragile than extending ArrayList itself.
Here’s our implementation of BoundedPriorityQueue following the same principle:

public class BoundedPriorityQueue<E> extends AbstractQueue<E> {
    private final Queue<E> elements;
    private final int bound;
    public BoundedPriorityQueue(Comparator<? super E> comparator, int bound) {
        this.elements = new PriorityQueue<>(comparator);
        this.bound = bound;
    }
    @Override public Iterator<E> iterator() { return elements.iterator(); }
    @Override public int size() { return elements.size(); }
    @Override public E poll() { return elements.poll(); }
    @Override public E peek() { return elements.peek(); }
    @Override
    public boolean offer(E e) {
        if (elements.size() > bound) return false;
        return elements.offer(e);
    }
}

Disadvantages

Of course, delegation has its disadvantages.

Memory overhead

There’s an additional object for the wrapper. This isn’t really a problem in practice, though.

Performance

You can’t take advantage of all optimizations a class may introduce in the future. ArrayList.removeIf is a great example again – the delegating class will only delegate a subset of methods, falling back to the potentially unoptimial behavior of AbstractList for others. But importantly, while this may be slower, it’s always correct – if it becomes a problem, just delegate that method as well. At least it won’t suddenly lead to broken behavior.

Missing Optional Methods

It’s possible to miss certain optional methods when delegating. In the standard AbstractList example, it’s very easy to forget to delegate the set(int, E) method since it isn’t used very often.
However, this is usually fail-fast – it’ll throw an Exception and you’ll notice it – and much preferrable to the unexpected and subtle results wrong extension may produce.

Additional Code

Especially with large interfaces and interfaces where you don’t have a handy stub class like AbstractList available, there can be quite a lot of methods to delegate. This is probably the biggest problem with delegation and a good reason to keep interfaces small.

Tooling

Your IDE can assist you with generating delegation methods.
If you are a fan of lombok, it offers an experimental @Delegate annotation. But beware: Using this annotation with only excludes defined can lead to the very same issue! Take the example from the lombok website:

class ExcludesDelegateExample {
  long counter = 0L;
  private interface Add {
    boolean add(String x);
    boolean addAll(Collection<? extends String> x);
  }
  @Delegate(excludes=Add.class)
  private final Collection<String> collection = new ArrayList<String>();
  public boolean add(String item) {
    counter++;
    return collection.add(item);
  }
  public boolean addAll(Collection<? extends String> col) {
    counter += col.size();
    return collection.addAll(col);
  }
}

Here we are missing proper handling for the listIterator.add and corresponding subList methods – we can enter invalid states and can have subtle bugs. Be very careful when using this annotation!

Conclusion

Don’t extend classes that weren’t meant to be extended. A class not being final does not give us a free pass – non-final is the default in java, and many programmers do not design their classes with extension in mind (and that’s fine).
Before extending a class with actual logic in it, think about alternatives. Can you use delegation instead? With collections and their convenient Abstract* stub implementations it’s easy, but with small interfaces you may not even need that.


JarSplice: building fat jars with native resources bundled

JarSplice is an application that builds a fat jar, with the main difference being that JarSplice will build invocation scripts and include native resources.
The main weaknesses of JarSplice seems to be its lack of build tool integration, its (apparently) unmaintained status, and its lack of source code availability (even though it’s described as being under the BSD license).
A “fat jar” (or a “shaded jar”) is a jar file that has been bundled with a project’s full set of required resources.
To explain, imagine that we have a Java project called “foo“, that depends on an external project, called “bar.”
For foo to run properly, then, our runtime classpath must contain foo.jar:bar.jar (in some order, although the order can be significant); both files must be present for the classloader to read in order for the project to run.
A “fat jar,” however, includes all of the resources from the classpath in a single jar. The build would extract every class from both foo.jar and bar.jar and built a new jar – named, for example, foo-all.jar – and therefore only this file would be required in order to run the project, since it has every resource that the prior classpath held.
However, this only works for Java resources; it doesn’t include native resources, like LWJGL‘s or SQLite‘s binary components. If foo uses SQLite, then the computer that foo is running on has to have SQLite deployed as a system library in order for foo to run.
This is what JarSplice addresses; it does bundle SQLite or LWJGL resources into the fat jar.
Again, the main weakness of JarSplice is that it’s not integrated into a build tool, so invoking it is an external mechanism for deployment. Development seems to have stopped (the home page says the code is BSD-licensed but no reference to the code exists.) Luis Quesada Torres built a command-line tool, JarSplicePlus; there are forks of that project that seem like there’s some promise for the future.
Feel free to participate and comment.

Getting a usable, non-localhost IP address for a JVM

I recently had a task where I needed to inform a docker image of the host OS’ IP address. (The docker image needed to make an HTTP call to the host machine, for testing purposes.) However, it’s not trivial to find code to actually get a host machine’s IP address. Here’s some code in Kotlin (and in Java) to accomplish the task; I can’t say it’s perfect (and won’t) but it worked for me, and hopefully this can serve as a tentpole for perhaps better code.
There are two pieces of code here; both eliminate the 10.*.*.* network. If you don’t need that network to be eliminated, well, remove that line.

Kotlin

import java.net.NetworkInterface
fun getAddress(): String {
    return NetworkInterface.getNetworkInterfaces().toList().flatMap {
        it.inetAddresses.toList()
                .filter { it.address.size == 4 }
                .filter { !it.isLoopbackAddress }
                .filter { it.address[0] != 10.toByte() }
                .map { it.hostAddress }
    }.first()
}

Java

The Java code’s a lot more convoluted, and isn’t helped at all by the requirement for a means by which to convert an Enumeration to a Stream (copied shamelessly from “Iterate an Enumeration in Java 8“, by the way.) However, it’s functionally equivalent to the Kotlin code. If it can’t find a valid address or an internal exception is thrown, a RuntimeException is created and thrown.

private static  Stream enumerationAsStream(Enumeration e) {
    return StreamSupport.stream(
            Spliterators.spliteratorUnknownSize(
                    new Iterator() {
                        public T next() {
                            return e.nextElement();
                        }
                        public boolean hasNext() {
                            return e.hasMoreElements();
                        }
                    },
                    Spliterator.ORDERED), false);
}
public static String getAddress() {
    try {
        Optional address = enumerationAsStream(NetworkInterface.getNetworkInterfaces())
                .flatMap(networkInterface ->
                        networkInterface.getInterfaceAddresses().stream()
                )
                .filter(ia -> !ia.getAddress().isLoopbackAddress())
                .map(InterfaceAddress::getAddress)
                .filter(a -> a.getAddress().length == 4)
                .filter(a -> a.getAddress()[0] != 10)
                .findFirst();
        if (address.isPresent()) {
            return address.get().getHostAddress();
        } else {
            throw new RuntimeException("Address not accessible");
        }
    } catch (SocketException e) {
        throw new RuntimeException(e);
    }
}

Summary

This is not perfect code, I wouldn’t think. It was thrown together to fit a specific need, and ideally should have been easily located on StackOverflow or something like that.

DropWizard Metrics Advice

I was working on an application with DropWizard, and I was having trouble getting my own metrics to show up in the display. The Metrics Getting Started is useful, and it actually showed me what I needed, but didn’t make it obvious enough for me.
What I needed was, in DropWizard Metrics parlance, a “meter.” This gives me performance data over time; basically, every time an event happens, I’d mark it and thus be able to see how busy the system was in the last minute, the last five minutes, and the last 15 minutes.
I followed the Metrics Getting Started:

  • I got a MetricsRegistry (by using new MetricsRegistry())
  • I created a Meter by calling register.meter(name) if necessary (and stored the Meter in a map so I could retrieve it again at will)
  • I marked an event by calling Meter.mark()

But at no point was I able to see the meter displayed in the DropWizard servlet.
The reason is because I created my own MetricsRegistry. One right way to do it is documented; it’s to use SharedMetricRegistries.getDefault(); instead (which gets you a MetricsRegistry that is displayed automatically).
Note that the DropWizard documentation is not wrong – it just steps past something that most people probably want by default.

LinkedList vs. ArrayList

Recently, The Java Programmer published “Difference between ArrayList and LinkedList in Java“, asserting different cases of when to prefer one List implementation over another. It’s a link full of conventional wisdom, and while it has some good information, it’s also wrong.
Here’s the channel’s factoid on LinkedList, as of 2017/Apr/25 (prior to it having been changed to point to the page you’re reading):

LinkedList is almost always slower than ArrayList/ArrayDeque. It might be best to use ArrayList/ArrayDeque to start. But if you have an actual performance problem, you can compare the two for your specific case. Otherwise, just use ArrayDeque. (It too has O(1) performance for insert-at-beginning, and unlike LinkedList is actually fast).

Implementation

The “Difference” page says that ArrayList uses an internal array to represent stored objects, while a LinkedList uses a doubly-linked list internally. In this, it’s correct.

Searching

The next point of comparison is entitled “Searching.” Here’s what it says about ArrayList:

An elements can be retrieved or accessed in O(1) time. ArrayList is faster because it uses array data structure and hence index based system is used to access elements.

And about LinkedList:

An element can be accessed in O(N) time. LinkedList is slower because it uses doubly linked list and for accessing an element it iterates from start or end (whichever is closer).

First off, in the interest of honesty, the last bit of information – that it iterates from either forward or backward, depending on which is closer – was a surprise to me (although it’s perfectly logical.) It’s also accurate.
However, this isn’t searching – it’s element access. Access and searching are different things; it’s get(42) vs. list.contains(matchingObject). This is important; contains() will have O(N) time for either List implementation, whereas ArrayList will have O(1) for get(), and LinkedList will have O(N) for get().
The thing is: LinkedList is not only O(N) for get() but the nature of the internal implementation makes it slower than one might think, because of cache locality. In an ArrayList, the element references are stored sequentially in memory; a LinkedList potentially splatters the references all over the heap. (This is not entirely likely, but it’s possible.) Thus, a LinkedList not only has O(N) performance for indexed access (where the ArrayList has O(1)), but the O(N) is worse than one might expect.
For accessing elements, there’s no situation apart from accessing the first or last element where a LinkedList competes with an ArrayList‘s speed.

Note that programming hates absolutes. It’s fully possible to create situations in which that last claim is not entirely true… but they’re rare and unlikely.

Insertion

About ArrayList:

Normally the insertion time is O(1). But in case array is full then array is resized by copying elements to new array, which makes insertion time O(N).

And about LinkedList:

Insertion of an element in LinkedList is faster, it takes O(1) time.

Incomplete data is provided here. For one thing, the author conflates mutation with “insertion.” There are two different kinds of additive list mutations (where data is added to a List): insertion (meaning that elements are prepended to other elements) and addition (where an element “follows” the last element in the list prior to mutation.)
Where the elements go as part of the additive mutation is really important, and it’s also important to note that the claim of LinkedList‘s O(1) time is very much a single type of insertion.
For an ArrayList, a list append (i.e., calling add(), which adds the provided element at the end of the list) has O(N) performance, but typically has O(1) performance. The O(N) complexity is because the internal array size might be exceeded by the addition of the element, in which case a new internal array is allocated, and the array references are copied over to the new internal array, and then the new element is appended. Thus, in the worst case, it is O(N), but this is misleading because:

  1. It’s fairly rare (the list resizes with a formula of currentSize*1.5 (the actual line is int newCapacity = oldCapacity + (oldCapacity >> 1);, if you’re interested.)
  2. Java’s blits are very, very, very fast; Java uses the blit mechanism constantly, and let’s be real, modern CPUs are really good at this anyway.

So is it actually O(N), then? … given what conventional Big O notation means, yes, it is; it’s just that O(N) is a lot less expensive than one might think in this case. And typically, the add() will be O(1) in actuality.
But what about LinkedList?
Insertion at the beginning or the end of the list is, in fact, O(1). This is even the common case (add(Object) calls linkLast(Object) by default.) However, if positional notation is used at all (i.e., add(int, Object) you degrade to O(N), because the LinkedList has to find the position at which the new element has to be inserted – and finding that position is O(N).
So what’s the conclusion? ArrayList does in fact have O(N) performance for adds, but it’s close enough to O(1) in real world conditions that we can colloquially speak of it being likely as fast as LinkedList‘s similar case… and if we add any kind of positional insertion, ArrayList‘s degradation under the worst case is far better than LinkedList‘s degradation.
As usual, we can create situations under which this is not the case (for example, if we always insert after the first position in a LinkedList, which will “search” one element and only that element) but these are fairly rare.

Deletion

Let’s see about ArrayList:

Deletion of an element is slower as all the elements have to be shifted to fill the space created by deleted element.

And LinkedList:

Deletion of an element is faster in LinkedList because no elements shifting are required.

Well, um, no.
Here, the author is confusing complexity – big O notation – with speed, for ArrayList. (He or she is wrong about LinkedList in any event.)
For an ArrayList, deleting an element by index is indeed O(N) because the JVM does have to blit the elements after the removed element (it has to shift them all in the internal array.)
However, for a LinkedList, it also has to find the elements around the element to remove – and if you recall, accessing a specific element for a LinkedList is an O(N) complexity itself. Even though the removal is simple (it’s simply relinking the nodes before and after the removed element), actually finding those nodes is O(N), thus deletion is O(N) for LinkedList – and slower than ArrayList, to boot.

Memory and Initial Capacity

The author says that an ArrayList has a default internal capacity of 10. This is incorrect for Java 8. See this code.
For the rest of it, the author is correct; a LinkedList is “empty” (there are no linked nodes internally); an ArrayList has capacity references in the internal list, while a LinkedList has a chain of objects, which will have higher memory consumption, but that’s not likely to matter unless your lists are huge.

The Comparison’s Conclusion

  • As search or get operation is faster in ArrayList so it is used in the scenario where more search or get operations are required.
  • As insertion or deletion of an element is faster in LinkedList so it is used in the scenario where more insertion and deletion operations are required.

The thing that the conclusion is missing is that “insertion and deletion” operations in LinkedList typically involve searches and gets – which are terrible for LinkedList – and therefore for even the operations that LinkedList is potentially better, ArrayList will typically perform better.
The factoid stands. If you need a List, use ArrayList first, then ArrayDeque if your access patterns demand… use LinkedList only after exhausting other possibilities.

Immutability in Java

Recently in ##java, we had a discussion on what constitutes “immutability” and how to implement it in java properly. Immutability is a core concept for functional programming, and as mainstream languages and ecosystems “catch up” with the safer, less error-prone ideas from functional programming, immutability is wanted in Java projects as well.

What is immutability?

Asking programmers for their definition of immutability, you will usually receive an answer such as this:

An object is immutable if its state never changes.

In this article we will look at different interpretations of that statement and see how to implement it in java.

Unmodifiable vs. Immutable

An unmodifiable object is an object that cannot be mutated. This does not mean it cannot change through other means:

List<String> inner = new ArrayList<>();
inner.add("foo");
List<String> unmodifiable = Collections.unmodifiableList(inner);
System.out.println(unmodifiable); // [foo]
inner.add("bar");
System.out.println(unmodifiable); // [foo, bar]

Unmodifiable objects are generally not considered “truly” immutable. An immutable version of the code above (using Guava‘s ImmutableList) would be:

List<String> inner = new ArrayList<>();
inner.add("foo");
List<String> immutable = ImmutableList.copyOf(inner);
System.out.println(immutable); // [foo]
inner.add("bar");
System.out.println(immutable); // [foo]

Polymorphism

Designing an immutable class to be extended is tricky business. The base class must either relax its immutability or make basically all of its methods final which makes polymorphism basically useless. Relaxing immutability guarantees will often lead to only an unmodifiable class:

public class Vec2 {
    private final int x;
    private final int y;
    public Vec2(int x, int y) {
        this.x = x;
        this.y = y;
    }
    public int getX() { return x; }
    public int getY() { return y; }
    @Override public int hashCode() { return getX() * 31 + getY(); }
    @Override public boolean equals(Object o) { ... }
}

This class appears immutable at first glance, but is it really?

public class MutableVec2 extends Vec2 {
    // redundant fields because super fields are private and final
    private int x;
    private int y;
    public MutableVec2(int x, int y) {
        super(x, y);
        this.x = x;
        this.y = y;
    }
    @Override public int getX() { return x; }
    @Override public int getY() { return y; }
    public void setX(int x) { this.x = x; }
    public void setY(int y) { this.y = y; }
}

We’ve suddenly introduced mutability, and not just that, we also override the two getters, so users of the original “immutable” Vec2 class cannot rely on instances of Vec2 never changing state when using them. This makes Vec2 effectively unmodifiable only, not immutable. Immutable classes should be final.
Limited polymorphism is possible when necessary, for example through a package-private constructor so immutability can be strictly controlled. This is error-prone and should be reserved for exceptional cases.

Defensive copies

Immutability itself does not require immutability on class members (in most definitions). However, in many cases where immutability is used, making sure that members also stay unmodified is desired. This can be done through “defensive copies”:

public class ImmutableClass {
    private final List<String> list;
    public ImmutableClass(List<String> list) {
        this.list = Collections.unmodifiableList(new ArrayList<>(list));
        // when using guava:
        //this.list = ImmutableList.copyOf(list);
    }
    // getter, equals, hashCode
}

Internal state

Internal mutable state is allowed in immutable classes. One use case of this is caching. In OpenJDKs java.lang.String class, which is immutable, the hashCode is cached using this method:

public class String {
    ...
    private int hash; // defaults to 0. not explicitly initialized to avoid race condition (see section below)
    @Override
    public int hashCode() {
        int h = hash;
        if (h == 0) {
            hash = h = computeHashCode();
        }
        return h;
    }
}

(This isn’t the actual implementation but it’s the same principle. Want to see this as implemented in an actual Java runtime library? Here it is.)
To users of this class, the internal state is never visible – they do not see a change in behavior. Because of this, the class still counts as immutable.

Concurrency

Java concurrency is a fickle beast, and immutability is an important tool to avoid the shared mutable state that makes concurrency so difficult. However, immutability has to be implemented carefully to work reliably in concurrent applications.
On top of the basic “state never changes” requirement we must introduce two more requirements: (from Java Concurrency In Practice. Read it, it’s a great book.)

  • All fields must be final.
  • The this reference must never escape the constructor. This is called “proper construction”.

To explore why these requirements are necessary, we will take an example straight from the Java Language Specification:

final class MaybeImmutable {
    private int x;
    public MaybeImmutable(int x) {
        this.x = x;
    }
    public int getX() { return x; }
    // the two threads
    static MaybeImmutable sharedField; // this field isn't volatile
    static void thread1() {
        sharedField = new MaybeImmutable(5);
    }
    static void thread2() {
        MaybeImmutable tmp = sharedField;
        if (tmp != null) {
            System.out.println(tmp.getX()); // could be 5... but could also be 0!
        }
    }
}

At first glance the MaybeImmutable class appears immutable – it does not provide any API to modify its state. The problem with this class is that the field x is not final. Let’s take a look at what both threads do:

// thread1
tmp = allocate MaybeImmutable
tmp.x = 5
write sharedField = tmp
// thread2
read tmp = sharedField
check tmp != null
print tmp.x

In this case, the constructor call is “just” setting the field. If you’re already familiar with the java memory model, you will notice that there is no happens-before ordering between those two threads! This makes the following execution order legal:

// thread1
tmp = allocate MaybeImmutable
write sharedField = tmp
tmp.x = 5
// thread2
// same as above

In that case, thread2 could legally print out “0”.
(Reordering is only one reason why this could happen – caches are another, for example.)
This fact makes our MaybeImmutable class not immutable: From thread2, the state of the same exact instance of MaybeImmutable can change over time. How can we fix this? By making the field x final. Looking at the JLS, this guarantees that when we can see the MyImmutable instance in the sharedField, we can also see the correctly initialized x field. thread2 will always print out 5, as we’d expect.
This visibility guarantee is only given if our class does not leak its this reference before construction is complete, which brings us to our second requirement (“proper construction”):

final class NotImmutable {
    public static NotImmutable instance;
    private final int x;
    public NotImmutable(int x) {
        this.x = x;
        instance = this;
    }
    public int getX() { return x; }
    static void thread1() {
        new NotImmutable(5);
    }
    static void thread2() {
        NotImmutable tmp = instance;
        if (tmp != null) {
            System.out.println(tmp.getX()); // could still be 0!
        }
    }
}

Even though the x field is final in this case, thread2 can see an instance of NotImmutable before its constructor completes, meaning it can still see 0 as the value of x.
Interestingly, setting the x field to volatile in doesn’t help in either example (NotImmutable or MaybeImmutable). The guarantees of final fields are actually different than those of volatile fields here. This is covered in Aleksey Shipilëv’s article “Close Encounters of The Java Memory Model Kind.” This article, together with “Java Memory Model Pragmatics,” is a great resource to learn more about the Java memory model.

Builders

Immutable classes, especially those with many fields, can be difficult to construct – the Java language lacks named parameters. This can be made somewhat easier through the builder pattern, which the following example demonstrates (though only with a single field):

public class ImmutableClass {
    private final String url;
    private ImmutableClass(String url) {
        this.url = url;
    }
    public String getX() {
        return url;
    }
    public static class Builder {
        private String url;
        public Builder url(String url) {
            this.url = url;
            return this;
        }
        public ImmutableClass build() {
            return new ImmutableClass(url);
        }
    }
}
ImmutableClass myInstance = new ImmutableClass.Builder().url("https://yawk.at/").build();

This pattern provides an easy way of constructing very large immutable classes – builders are infinitely more readable than constructor calls such as new ImmutableClass("GET", 6, true, true, false, 5, null). Most IDEs also include utilities to quickly generate such builders.

Lombok

The wonderful project lombok provides a very easy way of writing immutable classes in a few lines of code via the @Value annotation:

@Value
public class ImmutableClass {
    private final String url;
}

This makes the fields private final (if they aren’t already), generates getters and a constructor, and makes the class final. As an added bonus it also generates an equals and hashCode for you. On top of that, you can use the @Builder annotation to also let it generate a builder for you, and the experimental @Wither annotation to generate copy mutators.
To see what kind of code Lombok can generate you can either view the examples on their website or use my javap pastebin to see decompiled lombok-generated code (make sure to select “Procyon” on the bottom-right).

Serialization

Immutable classes do not follow the traditional Java bean layout: They do not have a zero-argument constructor and they do not have setters. This makes serialization using reflection difficult. There are solutions, though – an immutable class constructor can be annotated with serialization-framework-specific annotations to give the framework information on which parameter corresponds to which object property.
We will look at the jackson-databind annotations first:

public final class Vec2 {
    private final int x;
    private final int y;
    @JsonCreator
    public Vec2(@JsonProperty("x") int x, @JsonProperty("y") int y) {
        this.x = x;
        this.y = y;
    }
    public int getX() { return x; }
    public int getY() { return y; }
}

A more framework-agnostic solution exists using the java.beans.ConstructorProperties annotation (not present in android) which is supported by Jackson since version 2.7.0:

public final class Vec2 {
    private final int x;
    private final int y;
    @ConstructorProperties({ "x", "y" })
    public Vec2(int x, int y) {
        this.x = x;
        this.y = y;
    }
    public int getX() { return x; }
    public int getY() { return y; }
}

This annotation is also generated on Lombok constructors by default making serializing Lombok-generated immutable classes hassle-free with jackson 2.7+.
Gson takes another approach to deserialization of objects without a zero-argument constructor: sun.misc.Unsafe provides a method to allocate an instance of a class without calling its constructor. Gson then proceeds to reflectively set fields of the immutable object. However, Unsafe is unsupported API and the OpenJDK project is attempting to phase out the class and may remove it entirely in a future JDK version, so this solution should be avoided.
Immutability in Java is not as simple in Java as it might appear at first but with the right tools and patterns, it can be very beneficial to your code base, making your application easier to read, debug and extend and making parallelism much simpler. Be careful of accidentally adding mutability and make use of tools like Lombok to reduce boilerplate associated with immutable classes in Java.

Interesting Links – 2017-jan-23

  • Is it me, or is it actually ironic that the website for Rome (a utility library for RSS) doesn’t have an associated RSS feed?
  • The Log: What every software engineer should know about real-time data’s unifying abstraction, submitted by user whaley, is an excellently written article that discusses logs in the context of distributed systems – where the log is king, whether you use it or not. It’s a fantastic summary of distributed data processing from the standpoint of the technologies that the technique rely upon.
  • What Happens When You Mix Java with a 1960 IBM Mainframe which discusses a presentation by US Government employee Marianne Berlotti. In it, she describes Java having been used to serve as an integration layer between thoroughly modern technology and mainframes from the distant past (including providing a layer between JDBC and an IMS database – and for some reason, your author has a true fondness for heirarchical databases.) Neat stuff, even though the Java applications serve as performance problems in the architecture.

EJB Injection in JSP with Java EE 7

User suexec asked ##java about accessing an EJB from a JSP template.
User dreamreal (i.e., your author) put together a simple application (rather imaginatively called “suexec-war” ) to explore the possibilities. The application was built from Adam Bien‘s minimalistic Java EE 7 Maven archetype, deployed into Wildfly 10.0.
It turns out that to the best of my understanding, injection of an EJB into a JSP is simply unsupported. The JSP can use JNDI to acquire an EJB via JNDI (see source) but the CDI injection never worked.

Next, I wrote a simple servlet (called, of all things, TestServlet.java). Here, the @EJB MyEJB ejb worked like a charm.

To render, I used the Handlebars template library just for kicks; Freemarker (or even JSP itself) would have worked just as well, but I wanted to play with something different.
The process would have been the same no matter what templating library was chosen: you’d take the values you wanted to render and store them somewhere such that the rendering engine could access them. In my case, I could have built a composite object (or even asked Handlebars to call the greet() method itself, using the ejb value) but I was aiming for simplicity rather than optimal behavior, following whaley’s “Make it work, make it pretty, make it fast” principle (from “Suffering-oriented programming” )- this is rough code, meant to serve more as a smoke test than an actual example of a great application, so I left off at “make it work,” leaving even “make it pretty” to others.
Comments and improvements welcomed.