Develop Reference

Can anonymous classes be completely immutable?

In the book Java Concurrency In Practice, there is this example of an almost immutable object which is at risk of failure if not properly published:
// Taken from Java Concurrency In Practice
// p.51 Listing 3.15: Class at risk of failure if not properly published.
public class Holder {
private int n;
public Holder(int n) { this.n = n; }
public void assertSanity() {
if(n != n)
throw new AssertionError("This statement is false.");
}
}
// p.50 Listing 3.14: Publishing an object without adequate synchronization. Don't do this.
class Client {
public Holder holder;
public void initialize() {
holder = new Holder(42);
}
}
If I understand the chapter in the book correctly, adding final to the n field of the Holder class will make the object completely immutable and eliminate the chance of getting the AssertionError thrown even if it's still published without adequate synchronization like it's done in the Client class.
Now I'm wondering how anonymous classes behave in this respect. Please see the following example:
public interface IHolder {
void assertSanity();
}
class IHolderFactory {
static IHolder create(int n) {
return new IHolder() {
#Override
public void assertSanity() {
if (n != n)
throw new AssertionError("This statement is false.");
}
};
}
}
class IHolderClient {
public IHolder holder;
public void initialize() {
// is this safe?
holder = IHolderFactory.create(42);
}
}
It's published without adequate synchronization just like in the example from the book, but the difference is that now the Holder class has become an interface and there is a static factory method which returns an anonymous class implementing the interface, and the anonymous class uses the method parameter n.
My question is: is there any chance of getting the AssertionError from my latter example thrown? If there is, what is the best way to make it completely immutable and eliminate the problem? Does it change something if it was written in a functional way like the following?
class IHolderFactory {
static IHolder create(int n) {
return () -> {
if (n != n)
throw new AssertionError("This statement is false.");
};
}
}

This is a very tricky issue.
JLS, §17.4.1. Shared Variables says:
Local variables (§14.4), formal method parameters (§8.4.1), and exception handler parameters (§14.20) are never shared between threads and are unaffected by the memory model.
This seems to contradict the fact that you can use them within an inner class or lambda expression that can be shared between threads, but those constructs capture the value of the variable and use the value. This process, however, is not very well specified.
The only mentioning I could ever find, is in §15.27.2 explaining the (effective) final requirement:
The restriction to effectively final variables prohibits access to dynamically-changing local variables, whose capture would likely introduce concurrency problems.
In practice, the captured values are stored in synthetic final fields of the inner class or the class generated at runtime for the lambda expression. So you will never see the error with the current implementation.
This, however, is not specified anywhere. The language specification says little about the bytecode format and the virtual machine specification says little about the language constructs.
So, local variables, formal method parameters, and exception handler parameters are explicitly excluded from the JMM and their captured values are not variables in the JMM’s regard and not even mentioned there. The question is what does that mean.
Are they generally immune to data races (my interpretation) or are they unsafe and we do no get any guaranty from the JMM at all? In the latter case, it would even imply that we were not able to make them safe, as any safe publication mechanism gets its safety from guarantees of the JMM which does not cover our case. It’s worth noting that the JMM also does not cover the outer this reference nor an instance’s implicit reference to a Class object returned by getClass().
So while I’d consider them immune to data races, I wish that was specified more explicit.

It does not matter if you use an anonymous class or lambda, you have zero synchronization mechanisms here to correctly publish the reference; as such, this code can throw that Exception.
In other words, there are tools and conditions that you must meet so that your code is safe: these are using final, volatile or some sort of synchronized or locks, etc. Since you use none, no guarantees are provided.
The Java Language Specification offers these guarantees only when you use special semantics, like final that you have shown in the first example. There are others too, but making an object immutable is the simplest, most trivial way. This is the best article that I am aware of on this subject, you might want to read it.

Local static variables and Java

I was wondering how to achieve the local static variable in java. I know Java wount support it. But what is the better way to achieve the same? I donot want the other methods in my class to access the variable, but it should retain the value across the invocations of the method.
Can somebody please let me know.

I don't think there is any way to achieve this. Java does not support 'local static' a la C, and there is no way to retrofit this while still keeping your sourcecode "real Java"1.
I donot want the other methods in my class to access the variable, but it should retain the value across the invocations of the method.
The best thing would be to make it an ordinary (private) static, and then just don't access it from other methods. The last bit should be easy ... 'cos you are writing the class.
1 - I suppose you could hack something together that involves preprocessing your code, but that will make all sorts of other things unpleasant. My advice is don't go there: it is not worth the pain.

Rather than trying to actually protect the variable, making the code more obscure and complicated, consider logical protection by comment and placement. I declare normal fields at the start of the class, but a field that should only be accessed from one method just before that method. Include a comment saying it should only be used in the one method:
// i should be used only in f
private int i;
/**
* Documentation for f().
*/
public void f(){
System.out.println(i++);
}

What you want is the ability to constraint intermediate computation results within the relevant method itself. To achieve this, you can refer to the following code example. Suppose you want to maintain a static variable i across multiple calls of m(). Instead of having such a static variable, which is not feasible for Java, you can encapsulate variable i into a field of a class A visible only to m(), create a method f(), and move all your code for m() into f(). You can copy, compile, and run the following code, and see how it works.
public class S {
public void m() {
class A {
int i;
void f() {
System.out.println(i++);
}
}
A a = new A();
a.f();
a.f();
a.f();
}
public static void main(String[] args) {
S s = new S();
s.m();
}
}

In theory, yes - but not in conventional manners.
What I would do to create this:
Create that Object in a totally different class, under the private modifier, with no ability to be accessed directly.
Use a debugging tool, such as the JDI to find that variable in the other class, get it's ObjectReference and manipulate directly or create a new variable which references to that object, and use that variable, which references to the object, in your method.
This is quite complicated, as using the JDI is tough, and you would need to run your program on 2 processes.
If you want to do this, I suggest looking into the JDI, but my honest answer would be to look for another solution.

Based on dacongy's idea of using a method local class I created a simple solution:
public class Main {
public static String m() {
class Statics {
static String staticString;
}
if (Statics.staticString == null)
Statics.staticString = "My lazy static method local variable";
return Statics.staticString;
}
}

Java Variables Basics

Ok, so I am about to embarrass my self here but I am working on a project that I will need to get some help on so I need to get some conventions down so I don't look too stupid. I have only been doing java for 2 months and 100% of that has been on Android.
I need some help understanding setting up variables and why I should do it a certain way.
Here is an example of my variables list for a class:
Button listen,feed;
Context context = this;
int totalSize = 0;
int downloadedSize = 0;
SeekBar seek;
String[] feedContent = new String[1000];
String[] feedItems = new String[1000];
ListView podcast_list = null;
HtmlGrabber html = new HtmlGrabber();
String pkg = "com.TwitForAndroid";
TextView progress = null;
long cp = 0;
long tp = 0;
String source = null;
String pageContent = null;
String pageName = "http://www.shanescode.com";
DataBaseHelper mdbHelper = new DataBaseHelper(this);
int songdur = 0;
So all of these are variables that I want to use in all through the whole class. Why would I make something a static, or a final. I understand Public but why make something private?
Thanks for your help and please don't be too harsh. I just need some clarification.

These words all alter the way the variable to which they are applied can be used in code.
static means that the variable will only be created once for the entire class, rather than one for each different instance of that class.
public class MyClass{
public static int myNumber;
}
In this case the variable is accessed as MyClass.myNumber, rather than through an instance of MyClass. Static variables are used when you want to store data about the class as a whole rather than about an individual instance.
final prevents the variable's value from changing after it is set the first time. It must be given an initial value either as part of its declaration:
public final int myNumber = 3;
or as part of the class's constructor:
public MyClass(int number){
this.myNumber = 3;
Once this is done, the variable's value cannot be changed. Keep in mind, though, that if the variable is storing an object this does not prevent the object's variable from being changed. This keyword is used to keep a piece of data constant, which can make writing code using that data much easier.
private modifies the visibility of the variable. A private variable can be accessed by the instance which contains it, but not outside that:
public class MyClass{
private int myNumber;
public void changeNumber(int number){
this.myNumber = number; //this works
}
}
MyClass myInstance = new MyClass();
myInstance.myNumber = 3; //This does not work
myInstance.changeNumber(3) //This works
Visibility is used to control how a class's variables can be used by other code. This is very important when writing code which will be used by other programmers, in order to control how they can access the internal structure of your classes. Public and private are actually only two of the four possible levels of visibility in Java: the others are protected and "no (visibility) modifier" (a.k.a not public or private or protected). The differences between these four levels is detailed here.

static = same for all instances of a class.
final = unchanging (reference) for a particular instance.
If you needed some field (aka a class variable) to be shared by all instances of a class (e.g., a constant) then you might make it static.
If you know some field is immutable (at least, it's reference is immutable) in an instance, then it is good practice to make it final. Again, constants would be a good example of a field to make final; anything that is constant within an instance from construction time on is also a good candidate for final.
A search for "java final static" gives pretty useful further reference on the use of those keywords.
The use of the private keyword controls what can accessed by other classes. I'd say it's biggest use is to help developers "do the right thing" - instead of accessing the internals of the implementation of another class, which could produce all sorts of unwanted behavior, it forces using accessor/mutator methods, which the class implementor can use to enforce the appropriate constraints.

Private
The idea behind using private is information hiding. Forget about software for a second; imagine a piece of hardware, like an X-Box or something. Somewhere on it, it has a little hatch to access the inside, usually sporting a sticker: "open this up and warranty is void."
Using private is sticking a sticker like that in your software component; some things are 'inside' only, and while it would be easy for anyone to open it up and play with the inside anyways, you're letting them know that if they do, you're not responsible for the unexpected behavior that results.
Static
The static keyword does not mean "same for all instances of a class"; that's a simplification. Rather, it is the antonym of "dynamic". Using the static keyword means "There is no dynamic dispatching on this member." This means that the compiler and not the run-time determines what code executes when you call this method.
Since thee are no instances of objects at compile-time this means that a static member has no access to an instance.
An example:
public class Cat {
public static void speak() { System.out.println("meow"); }
}
public class Lion extends Cat {
public static void speak() { System.out.println("ROAR"); }
}
// ...
public static void main(String argv[]) {
Cat c = new Lion();
c.speak();
}
The above prints "meow" - not "roar" - because speak is a static member, and the declared type of c is Cat, so the compiler builds in such a way that Cat.speak is executed, not Lion.speak. Were there dynamic dispatching on static members, then Lion.speak would execute, as the run-time type of c is Lion.

Another thing that might trip you up is this:
Not everything has to be a class level variable; you should have a variable defined for the smallest scope it needs to be defined.
So as an example, suppose your class only has one method which uses your TextView progress variable. Move that declaration into the method that needs it. This way it tidies things up and helps you make more robust code by separating out things that are really separate.

I don't know why you would make anything private.
Folks will chime in and say that private is a Very Important Thing.
Some folks will claim that you can't do encapsulation without private. Most of this seems to be privacy for privacy's sake.
If you are selling your code to someone else, then you must carefully separate the interface elements of your class from the implementation details of your class. In this case, you want to make the implementation private (or protected) so that -- for legal purposes -- the code you sell doesn't expose too much of the implementation details.
Otherwise, if you're not selling it, don't waste a lot of time on private.
Invest your time in separating Interface from Implementation. Document the Interface portions carefully to be sure you're playing by the rules. Clearly and cleanly keep the implementation details separate. Consider using private as a way to have the compiler "look over your shoulder" to be sure you've really separated interface from implementation.

One of the aspects of the object oriented approach that has made it so wildly popular is that you can hide your variables inside of a class. The class becomes like a container. Now you as the programmer get to decide how you want the users of your class to interact with it. In Java, the tradition is to provide an API -- a public interface for your class using methods of the class.
To make this approach work, you declare your variables as private ( which means only methods within your class can access them ) and then provide other methods to access them. For example,
private int someNumber;
This variable can only be accessed from within your class. Do you think others might need access to it from outside of the class? You would create a method to allow access:
public int getSomeNumber()
{
return someNumber;
}
Perhaps users of your class will also need the ability to set someNumber as well. In that case, you provide a method to do that as well:
public void setSomeNumber( int someNumber )
{
this.someNumber = someNumber;
}
Why all of this work just to get access to a class member that you could just as easily declare as public? If you do it using this approach, you have control over how others access the data in your class. Imagine that you want to make sure that someNumber only gets set to be a number < 100. You can provide that check in your setSomeNumber method. By declaring your variables to have private access, you protect your class from getting used incorrectly, and make it easier on everyone who needs to use it -- including yourself!
Declaring a variable to have static access means that you do not need an instance of the class to access the variable. In Java, generally you write a class and then create an instance of it. You can have as many instances of that class as you want, and they all keep track of their own data. You can also declare variables that are part of the class itself, and this is where the static keyword comes in. If you create a variable...
static int classVariable = 0;
the variable can be accessed without a class instance. For example, you might see this done from time to time:
public static final int MY_CONSTANT = 1;
While there are better ways to do this now, it is still a common pattern. You use this variable without any instance of the class like this:
myInstance.setSomeNumber( MyClass.MY_CONSTANT );
java.awt.Color uses static variables this way. You can also declare methods to be static ( look at public static void main, the starting point for your programs ). Statics are useful, but use them sparingly because creating instances of classes can often result in better designs.
Finally ( pun intended ), why would you ever want to declare a variable to be final? If you know that the value should never change, declaring it as final means that if you write some code that tries to change that value, the compiler will start complaining. This again helps protect from making silly mistakes that can add up to really annoying bugs.
If you look at the static variable example above, the final keyword is also used. This is a time when you have decided that you want to make a variable public, but also want to protect it from being changed. You do this by making it public and final.

Why should I use the keyword "final" on a method parameter in Java?

I can't understand where the final keyword is really handy when it is used on method parameters.
If we exclude the usage of anonymous classes, readability and intent declaration then it seems almost worthless to me.
Enforcing that some data remains constant is not as strong as it seems.
If the parameter is a primitive then it will have no effect since the parameter is passed to the method as a value and changing it will have no effect outside the scope.
If we are passing a parameter by reference, then the reference itself is a local variable and if the reference is changed from within the method, that would not have any effect from outside of the method scope.
Consider the simple test example below.
This test passes although the method changed the value of the reference given to it, it has no effect.
public void testNullify() {
Collection<Integer> c = new ArrayList<Integer>();
nullify(c);
assertNotNull(c);
final Collection<Integer> c1 = c;
assertTrue(c1.equals(c));
change(c);
assertTrue(c1.equals(c));
}
private void change(Collection<Integer> c) {
c = new ArrayList<Integer>();
}
public void nullify(Collection<?> t) {
t = null;
}

Stop a Variable’s Reassignment
While these answers are intellectually interesting, I've not read the short simple answer:
Use the keyword final when you want the compiler to prevent a
variable from being re-assigned to a different object.
Whether the variable is a static variable, member variable, local variable, or argument/parameter variable, the effect is entirely the same.
Example
Let’s see the effect in action.
Consider this simple method, where the two variables (arg and x) can both be re-assigned different objects.
// Example use of this method:
// this.doSomething( "tiger" );
void doSomething( String arg ) {
String x = arg; // Both variables now point to the same String object.
x = "elephant"; // This variable now points to a different String object.
arg = "giraffe"; // Ditto. Now neither variable points to the original passed String.
}
Mark the local variable as final. This results in a compiler error.
void doSomething( String arg ) {
final String x = arg; // Mark variable as 'final'.
x = "elephant"; // Compiler error: The final local variable x cannot be assigned.
arg = "giraffe";
}
Instead, let’s mark the parameter variable as final. This too results in a compiler error.
void doSomething( final String arg ) { // Mark argument as 'final'.
String x = arg;
x = "elephant";
arg = "giraffe"; // Compiler error: The passed argument variable arg cannot be re-assigned to another object.
}
Moral of the story:
If you want to ensure a variable always points to the same object,
mark the variable final.
Never Reassign Arguments
As good programming practice (in any language), you should never re-assign a parameter/argument variable to an object other than the object passed by the calling method. In the examples above, one should never write the line arg = . Since humans make mistakes, and programmers are human, let’s ask the compiler to assist us. Mark every parameter/argument variable as 'final' so that the compiler may find and flag any such re-assignments.
In Retrospect
As noted in other answers…
Given Java's original design goal of helping programmers to avoid dumb mistakes such as reading past the end of an array, Java should have been designed to automatically enforce all parameter/argument variables as 'final'. In other words, Arguments should not be variables. But hindsight is 20/20 vision, and the Java designers had their hands full at the time.
So, always add final to all arguments?
Should we add final to each and every method parameter being declared?
In theory, yes.
In practice, no.➥ Add final only when the method’s code is long or complicated, where the argument may be mistaken for a local or member variable and possibly re-assigned.
If you buy into the practice of never re-assigning an argument, you will be inclined to add a final to each. But this is tedious and makes the declaration a bit harder to read.
For short simple code where the argument is obviously an argument, and not a local variable nor a member variable, I do not bother adding the final. If the code is quite obvious, with no chance of me nor any other programmer doing maintenance or refactoring accidentally mistaking the argument variable as something other than an argument, then don’t bother. In my own work, I add final only in longer or more involved code where an argument might mistaken for a local or member variable.
#Another case added for the completeness
public class MyClass {
private int x;
//getters and setters
}
void doSomething( final MyClass arg ) { // Mark argument as 'final'.
arg = new MyClass(); // Compiler error: The passed argument variable arg cannot be re-assigned to another object.
arg.setX(20); // allowed
// We can re-assign properties of argument which is marked as final
}
record
Java 16 brings the new records feature. A record is a very brief way to define a class whose central purpose is to merely carry data, immutably and transparently.
You simply declare the class name along with the names and types of its member fields. The compiler implicitly provides the constructor, getters, equals & hashCode, and toString.
The fields are read-only, with no setters. So a record is one case where there is no need to mark the arguments final. They are already effectively final. Indeed, the compiler forbids using final when declaring the fields of a record.
public record Employee( String name , LocalDate whenHired ) // 🡄 Marking `final` here is *not* allowed.
{
}
If you provide an optional constructor, there you can mark final.
public record Employee(String name , LocalDate whenHired) // 🡄 Marking `final` here is *not* allowed.
{
public Employee ( final String name , final LocalDate whenHired ) // 🡄 Marking `final` here *is* allowed.
{
this.name = name;
whenHired = LocalDate.MIN; // 🡄 Compiler error, because of `final`.
this.whenHired = whenHired;
}
}

Sometimes it's nice to be explicit (for readability) that the variable doesn't change. Here's a simple example where using final can save some possible headaches:
public void setTest(String test) {
test = test;
}
If you forget the 'this' keyword on a setter, then the variable you want to set doesn't get set. However, if you used the final keyword on the parameter, then the bug would be caught at compile time.

Yes, excluding anonymous classes, readability and intent declaration it's almost worthless. Are those three things worthless though?
Personally I tend not to use final for local variables and parameters unless I'm using the variable in an anonymous inner class, but I can certainly see the point of those who want to make it clear that the parameter value itself won't change (even if the object it refers to changes its contents). For those who find that adds to readability, I think it's an entirely reasonable thing to do.
Your point would be more important if anyone were actually claiming that it did keep data constant in a way that it doesn't - but I can't remember seeing any such claims. Are you suggesting there's a significant body of developers suggesting that final has more effect than it really does?
EDIT: I should really have summed all of this up with a Monty Python reference; the question seems somewhat similar to asking "What have the Romans ever done for us?"

Let me explain a bit about the one case where you have to use final, which Jon already mentioned:
If you create an anonymous inner class in your method and use a local variable (such as a method parameter) inside that class, then the compiler forces you to make the parameter final:
public Iterator<Integer> createIntegerIterator(final int from, final int to)
{
return new Iterator<Integer>(){
int index = from;
public Integer next()
{
return index++;
}
public boolean hasNext()
{
return index <= to;
}
// remove method omitted
};
}
Here the from and to parameters need to be final so they can be used inside the anonymous class.
The reason for that requirement is this: Local variables live on the stack, therefore they exist only while the method is executed. However, the anonymous class instance is returned from the method, so it may live for much longer. You can't preserve the stack, because it is needed for subsequent method calls.
So what Java does instead is to put copies of those local variables as hidden instance variables into the anonymous class (you can see them if you examine the byte code). But if they were not final, one might expect the anonymous class and the method seeing changes the other one makes to the variable. In order to maintain the illusion that there is only one variable rather than two copies, it has to be final.

I use final all the time on parameters.
Does it add that much? Not really.
Would I turn it off? No.
The reason: I found 3 bugs where people had written sloppy code and failed to set a member variable in accessors. All bugs proved difficult to find.
I'd like to see this made the default in a future version of Java. The pass by value/reference thing trips up an awful lot of junior programmers.
One more thing.. my methods tend to have a low number of parameters so the extra text on a method declaration isn't an issue.

Using final in a method parameter has nothing to do with what happens to the argument on the caller side. It is only meant to mark it as not changing inside that method. As I try to adopt a more functional programming style, I kind of see the value in that.

Personally I don't use final on method parameters, because it adds too much clutter to parameter lists.
I prefer to enforce that method parameters are not changed through something like Checkstyle.
For local variables I use final whenever possible, I even let Eclipse do that automatically in my setup for personal projects.
I would certainly like something stronger like C/C++ const.

Since Java passes copies of arguments I feel the relevance of final is rather limited. I guess the habit comes from the C++ era where you could prohibit reference content from being changed by doing a const char const *. I feel this kind of stuff makes you believe the developer is inherently stupid as f*** and needs to be protected against truly every character he types. In all humbleness may I say, I write very few bugs even though I omit final (unless I don't want someone to override my methods and classes). Maybe I'm just an old-school dev.

Short answer: final helps a tiny bit but... use defensive programming on the client side instead.
Indeed, the problem with final is that it only enforces the reference is unchanged, gleefully allowing the referenced object members to be mutated, unbeknownst to the caller. Hence the best practice in this regard is defensive programming on the caller side, creating deeply immutable instances or deep copies of objects that are in danger of being mugged by unscrupulous APIs.

I never use final in a parameter list, it just adds clutter like previous respondents have said. Also in Eclipse you can set parameter assignment to generate an error so using final in a parameter list seems pretty redundant to me.
Interestingly when I enabled the Eclipse setting for parameter assignment generating an error on it caught this code (this is just how I remember the flow, not the actual code. ) :-
private String getString(String A, int i, String B, String C)
{
if (i > 0)
A += B;
if (i > 100)
A += C;
return A;
}
Playing devil's advocate, what exactly is wrong with doing this?

One additional reason to add final to parameter declarations is that it helps to identify variables that need to be renamed as part of a "Extract Method" refactoring. I have found that adding final to each parameter prior to starting a large method refactoring quickly tells me if there are any issues I need to address before continuing.
However, I generally remove them as superfluous at the end of the refactoring.

Follow up by Michel's post. I made myself another example to explain it. I hope it could help.
public static void main(String[] args){
MyParam myParam = thisIsWhy(new MyObj());
myParam.setArgNewName();
System.out.println(myParam.showObjName());
}
public static MyParam thisIsWhy(final MyObj obj){
MyParam myParam = new MyParam() {
#Override
public void setArgNewName() {
obj.name = "afterSet";
}
#Override
public String showObjName(){
return obj.name;
}
};
return myParam;
}
public static class MyObj{
String name = "beforeSet";
public MyObj() {
}
}
public abstract static class MyParam{
public abstract void setArgNewName();
public abstract String showObjName();
}
From the code above, in the method thisIsWhy(), we actually didn't assign the [argument MyObj obj] to a real reference in MyParam. In instead, we just use the [argument MyObj obj] in the method inside MyParam.
But after we finish the method thisIsWhy(), should the argument(object) MyObj still exist?
Seems like it should, because we can see in main we still call the method showObjName() and it needs to reach obj. MyParam will still use/reaches the method argument even the method already returned!
How Java really achieve this is to generate a copy also is a hidden reference of the argument MyObj obj inside the MyParam object ( but it's not a formal field in MyParam so that we can't see it )
As we call "showObjName", it will use that reference to get the corresponding value.
But if we didn't put the argument final, which leads a situation we can reassign a new memory(object) to the argument MyObj obj.
Technically there's no clash at all! If we are allowed to do that, below will be the situation:
We now have a hidden [MyObj obj] point to a [Memory A in heap] now live in MyParam object.
We also have another [MyObj obj] which is the argument point to a [Memory B in heap] now live in thisIsWhy method.
No clash, but "CONFUSING!!" Because they are all using the same "reference name" which is "obj".
To avoid this, set it as "final" to avoid programmer do the "mistake-prone" code.

Why must delegation to a different constructor happen first in a Java constructor?

In a constructor in Java, if you want to call another constructor (or a super constructor), it has to be the first line in the constructor. I assume this is because you shouldn't be allowed to modify any instance variables before the other constructor runs. But why can't you have statements before the constructor delegation, in order to compute the complex value to the other function? I can't think of any good reason, and I have hit some real cases where I have written some ugly code to get around this limitation.
So I'm just wondering:
Is there a good reason for this limitation?
Are there any plans to allow this in future Java releases? (Or has Sun definitively said this is not going to happen?)
For an example of what I'm talking about, consider some code I wrote which I gave in this StackOverflow answer. In that code, I have a BigFraction class, which has a BigInteger numerator and a BigInteger denominator. The "canonical" constructor is the BigFraction(BigInteger numerator, BigInteger denominator) form. For all the other constructors, I just convert the input parameters to BigIntegers, and call the "canonical" constructor, because I don't want to duplicate all the work.
In some cases this is easy; for example, the constructor that takes two longs is trivial:
public BigFraction(long numerator, long denominator)
{
this(BigInteger.valueOf(numerator), BigInteger.valueOf(denominator));
}
But in other cases, it is more difficult. Consider the constructor which takes a BigDecimal:
public BigFraction(BigDecimal d)
{
this(d.scale() < 0 ? d.unscaledValue().multiply(BigInteger.TEN.pow(-d.scale())) : d.unscaledValue(),
d.scale() < 0 ? BigInteger.ONE : BigInteger.TEN.pow(d.scale()));
}
I find this pretty ugly, but it helps me avoid duplicating code. The following is what I'd like to do, but it is illegal in Java:
public BigFraction(BigDecimal d)
{
BigInteger numerator = null;
BigInteger denominator = null;
if(d.scale() < 0)
{
numerator = d.unscaledValue().multiply(BigInteger.TEN.pow(-d.scale()));
denominator = BigInteger.ONE;
}
else
{
numerator = d.unscaledValue();
denominator = BigInteger.TEN.pow(d.scale());
}
this(numerator, denominator);
}
Update
There have been good answers, but thus far, no answers have been provided that I'm completely satisfied with, but I don't care enough to start a bounty, so I'm answering my own question (mainly to get rid of that annoying "have you considered marking an accepted answer" message).
Workarounds that have been suggested are:
Static factory.
I've used the class in a lot of places, so that code would break if I suddenly got rid of the public constructors and went with valueOf() functions.
It feels like a workaround to a limitation. I wouldn't get any other benefits of a factory because this cannot be subclassed and because common values are not being cached/interned.
Private static "constructor helper" methods.
This leads to lots of code bloat.
The code gets ugly because in some cases I really need to compute both numerator and denominator at the same time, and I can't return multiple values unless I return a BigInteger[] or some kind of private inner class.
The main argument against this functionality is that the compiler would have to check that you didn't use any instance variables or methods before calling the superconstructor, because the object would be in an invalid state. I agree, but I think this would be an easier check than the one which makes sure all final instance variables are always initialized in every constructor, no matter what path through the code is taken. The other argument is that you simply can't execute code beforehand, but this is clearly false because the code to compute the parameters to the superconstructor is getting executed somewhere, so it must be allowed at a bytecode level.
Now, what I'd like to see, is some good reason why the compiler couldn't let me take this code:
public MyClass(String s) {
this(Integer.parseInt(s));
}
public MyClass(int i) {
this.i = i;
}
And rewrite it like this (the bytecode would be basically identical, I'd think):
public MyClass(String s) {
int tmp = Integer.parseInt(s);
this(tmp);
}
public MyClass(int i) {
this.i = i;
}
The only real difference I see between those two examples is that the "tmp" variable's scope allows it to be accessed after calling this(tmp) in the second example. So maybe a special syntax (similar to static{} blocks for class initialization) would need to be introduced:
public MyClass(String s) {
//"init{}" is a hypothetical syntax where there is no access to instance
//variables/methods, and which must end with a call to another constructor
//(using either "this(...)" or "super(...)")
init {
int tmp = Integer.parseInt(s);
this(tmp);
}
}
public MyClass(int i) {
this.i = i;
}

I think several of the answers here are wrong because they assume encapsulation is somehow broken when calling super() after invoking some code. The fact is that the super can actually break encapsulation itself, because Java allows overriding methods in the constructor.
Consider these classes:
class A {
protected int i;
public void print() { System.out.println("Hello"); }
public A() { i = 13; print(); }
}
class B extends A {
private String msg;
public void print() { System.out.println(msg); }
public B(String msg) { super(); this.msg = msg; }
}
If you do
new B("Wubba lubba dub dub");
the message printed out is "null". That's because the constructor from A is accessing the uninitialized field from B. So frankly it seems that if someone wanted to do this:
class C extends A {
public C() {
System.out.println(i); // i not yet initialized
super();
}
}
Then that's just as much their problem as if they make class B above. In both cases the programmer has to know how the variables are accessed during construction. And given that you can call super() or this() with all kinds of expressions in the parameter list, it seems like an artificial restriction that you can't compute any expressions before calling the other constructor. Not to mention that the restriction applies to both super() and this() when presumably you know how to not break your own encapsulation when calling this().
My verdict: This feature is a bug in the compiler, perhaps originally motivated by a good reason, but in its current form it is an artifical limitation with no purpose.

I find this pretty ugly, but it helps
me avoid duplicating code. The
following is what I'd like to do, but
it is illegal in Java ...
You could also work around this limitation by using a static factory method that returns a new object:
public static BigFraction valueOf(BigDecimal d)
{
// computate numerator and denominator from d
return new BigFraction(numerator, denominator);
}
Alternatively, you could cheat by calling a private static method to do the computations for your constructor:
public BigFraction(BigDecimal d)
{
this(computeNumerator(d), computeDenominator(d));
}
private static BigInteger computeNumerator(BigDecimal d) { ... }
private static BigInteger computeDenominator(BigDecimal d) { ... }

The constructors must be called in order, from the root parent class to the most derived class. You can't execute any code beforehand in the derived constructor because before the parent constructor is called, the stack frame for the derived constructor hasn't even been allocated yet, because the derived constructor hasn't started executing. Admittedly, the syntax for Java doesn't make this fact clear.
Edit: To summarize, when a derived class constructor is "executing" before the this() call, the following points apply.
Member variables can't be touched, because they are invalid before base
classes are constructed.
Arguments are read-only, because the stack frame has not been allocated.
Local variables cannot be accessed, because the stack frame has not been allocated.
You can gain access to arguments and local variables if you allocated the constructors' stack frames in reverse order, from derived classes to base classes, but this would require all frames to be active at the same time, wasting memory for every object construction to allow for the rare case of code that wants to touch local variables before base classes are constructed.

"My guess is that, until a constructor has been called for every level of the heierarchy, the object is in an invalid state. It is unsafe for the JVM to run anything on it until it has been completely constructed."
Actually, it is possible to construct objects in Java without calling every constructor in the hierarchy, although not with the new keyword.
For example, when Java's serialization constructs an object during deserialization, it calls the constructor of the first non-serializable class in the hierarchy. So when java.util.HashMap is deserialized, first a java.util.HashMap instance is allocated and then the constructor of its first non-serializable superclass java.util.AbstractMap is called (which in turn calls java.lang.Object's constructor).
You can also use the Objenesis library to instantiate objects without calling the constructor.
Or if you are so inclined, you can generate the bytecode yourself (with ASM or similar). At the bytecode level, new Foo() compiles to two instructions:
NEW Foo
INVOKESPECIAL Foo.<init> ()V
If you want to avoid calling the constructor of Foo, you can change the second command, for example:
NEW Foo
INVOKESPECIAL java/lang/Object.<init> ()V
But even then, the constructor of Foo must contain a call to its superclass. Otherwise the JVM's class loader will throw an exception when loading the class, complaining that there is no call to super().

Allowing code to not call the super constructor first breaks encapsulation - the idea that you can write code and be able to prove that no matter what someone else does - extend it, invoke it, instansiate it - it will always be in a valid state.
IOW: it's not a JVM requirement as such, but a Comp Sci requirement. And an important one.
To solve your problem, incidentally, you make use of private static methods - they don't depend on any instance:
public BigFraction(BigDecimal d)
{
this(appropriateInitializationNumeratorFor(d),
appropriateInitializationDenominatorFor(d));
}
private static appropriateInitializationNumeratorFor(BigDecimal d)
{
if(d.scale() < 0)
{
return d.unscaledValue().multiply(BigInteger.TEN.pow(-d.scale()));
}
else
{
return d.unscaledValue();
}
}
If you don't like having separate methods (a lot of common logic you only want to execute once, for instance), have one method that returns a private little static inner class which is used to invoke a private constructor.

My guess is that, until a constructor has been called for every level of the heierarchy, the object is in an invalid state. It is unsafe for the JVM to run anything on it until it has been completely constructed.

Well, the problem is java cannot detect what 'statements' you are going to put before the super call. For example, you could refer to member variables which are not yet initialized. So I don't think java will ever support this.
Now, there are many ways to work around this problem such as by using factory or template methods.

Look it this way.
Let's say that an object is composed of 10 parts.
1,2,3,4,5,6,7,8,9,10
Ok?
From 1 to 9 are in the super class, part #10 is your addition.
Simple cannot add the 10th part until the previous 9 are completed.
That's it.
If from 1-6 are from another super class that fine, the thing is one single object is created in a specific sequence, that's the way is was designed.
Of course real reason is far more complex than this, but I think this would pretty much answers the question.
As for the alternatives, I think there are plenty already posted here.